This paper examines the manner in which novel forms of humanistic knowledge emerge in contexts of digital transformation through the computational analysis of the complete digitized work in three volumes by Colombian writer Tomás Carrasquilla. In dialogue with the emergence/ia theme of ACH 2026, the study addresses Latin American literary digitization as a simultaneous response to a double emergency: the technological expansion of artificial intelligence and the persistent underrepresentation of Spanish-speaking corpora within the global infrastructures of the digital humanities. From this perspective, the digital environment is conceptualized as a space for critical creation where cultural memory, computational analysis, and epistemic justice converge. The project's objective is to develop a hybrid methodology that integrates corpus linguistics and language models. This methodology articulates close philological reading with quantitative analysis and algorithmic modeling. In the initial phase, corpus tools (AntConc and LancsBox) were employed, encompassing frequency lists, concordances (KWIC), collocation analysis (MI and t-score), identification of n-grams, and visualization of lexical dispersion between subcorpora. The findings indicate that the lexical configurations observed are associated with rural Antioquia, quotidian religiosity, kinship networks, and traditional economies. This suggests that regional literature functions as a long-term linguistic and cultural archive in the Americas. In the subsequent phase, generative language models (ChatGPT and Gemini) independently extracted a lexicon that was characteristic of the corpus. Despite their concurrence within predominant semantic domains, they concomitantly engendered lexical hallucinations, accelerated modernization processes, and diminished sensitivity to diatopic variations. The contrast between statistical analysis and algorithmic synthesis underscores the interpretive potential of AI and its epistemological limits in multilingual and regional contexts. The study proposes cross-validation between methods as a critical practice in Latin American digital humanities, capable of contributing to the creation of more inclusive knowledge infrastructures. Consequently, the project exemplifies the manner in which the digital analysis of regional literary archives not only broadens the computational study of literature but also engenders situated models of knowledge production that respond inventively to contemporary technological emergencies.

The hype around AI at all levels of the university represents a multi-faceted emergenc(e/y) in the humanities. While generative AI poses immediate challenges to traditional pedagogical approaches in spaces like rhetoric and composition, other types of AI tools – such as carefully-tuned models for particular kinds of historic handwriting – offer a great deal of promise for meaningfully facilitating both teaching and research. The Transkribus platform, initially developed with European Union funding, has made considerable strides towards offering an environment that does effective transcription, model fine-tuning, and user-driven text correction, without being offputting to less-technical users (Nockels et al 2022). The user-friendliness is an important consideration: the scholars with the greatest linguistic expertise, particularly for smaller historical languages, are typically older, and have dedicated their careers to deep engagement with these languages and cultures. While they are curious and engaged with technology, an unfamiliar interface (such as the command line, or even a Jupyter notebook) is a significant barrier to adopting a tool. Using Transkribus, we are developing the first open-source AI transcription model for Older (Classical) Nahuatl, designed to support both scholarship and public learning through direct engagement with colonial Nahua documents. Despite decades of work, Nahuatl research still faces persistent, foundational questions about the meaning and function of key particles and morphemes—questions that remain difficult to resolve because Nahuatl is agglutinative and evidence is dispersed across thousands of manuscript pages. In our first phase, we are producing diplomatic transcriptions of a large archive of classical Nahuatl manuscripts collected by John Sullivan (University of Warsaw). These transcriptions will form a high-quality training dataset and lay the foundation for a second phase of annotating the transcriptions with information about particles and morphemes, in order to train an AI model to predict this information for additional texts. There is a substantial international community—including researchers at Yale, Princeton, Harvard, Tulane, Northwestern, and UNAM—who would immediately benefit from a shared corpus and model, which would enable faster and more reliable reading of difficult documents, improved lexical and grammatical research, and new kinds of questions that can only be answered with large-scale evidence (for example, shifts in particle usage across time, scribal habits, regional variation, and genre-specific conventions). At the same time, the project will open access to colonial Nahua sources for students and non-specialists. In this talk we will cover progress on adapting Transkribus models (such as in Murrieta-Flores 2025) for transcribing colonial Spanish into ones that can perform well with Nahuatl, and creating a corpus that can be annotated by experts. This can serve as training data for future transcription algorithms that offer greater accuracy than Transkribus’s PyLaia. References: Murrieta-Flores, Patricia. “Unlocking colonial records with Artificial Intelligence. Achieving the automated transcription of large-scale 16th and 17th-century Latin American historical collections”. STAR: Science & Technology of Archaeological Research, vol. 11, 2025. Nockels, Joe et al. “Understanding the application of handwritten text recognition technology in heritage contexts: a systematic review of Transkribus in published research”. Archival Science, vol. 22, 2022.

This paper presents a community-driven, collaborative project working toward the development of a text-to-speech (TTS) system for the Passamaquoddy language, a severely endangered Algonquian language spoken in the northeastern Americas. In the spirit of ACH 2026's Emergence/ia theme, this research asks: How can digital humanities methodologies foster the emergence of language technologies that honor community sovereignty with care, accountability, and transformative possibility? The project is grounded in authentic partnership with Passamaquoddy elders, language teachers, and tribal leadership. The research integrates Western and Indigenous epistemologies to develop a methodology that intertwines the conceptual Glocademia Matrix framework (Guilherme,2022) and the spiderweb conceptual framework (Enos, 20017) to explain this partnership. This approach reflects the project's core commitment that language technology must serve community-defined revitalization goals, not reproduce the extractive logics that have long endangered Indigenous languages and knowledge systems (Meighan, 2021). Data collection employs semi-structured interviews with tribal members to understand community desires for expanding language use, with responses analyzed to surface key themes. Alongside this, the project gathers text-based materials and voice recordings from fluent speakers using SpeechRecorder and Audacity, building a foundational corpus for prototype TTS development. Together, these methods center community voices at every stage of technological design. Key findings illuminate the linguistic and cultural considerations essential for building a TTS system that aligns with Passamaquoddy values—including how the technology can maintain tribal sovereignty over language resources and avoid digital colonialism (Ní Chasaide et al., 2023; Schwartz, Schneider & Chen, 2019). The paper addresses how community-led approaches to language technology can balance innovation with cultural preservation, and how DH scholars working alongside Indigenous communities must navigate questions of power, consent, and the long-term stewardship of language data. This work contributes to conversations at the intersection of digital humanities, computational creativity, multilingualism, and resource creation. The project offers a model of transnational solidarity and responsiveness that foregrounds the voices of those most affected by both linguistic emergency and technological emergence.

Computational film analysis confronts a crisis. Its most powerful methods advance scholarship, but at substantial environmental and accessibility costs. As Wevers and Smits (2020) argue, convolutional neural networks now enable corpus-level analysis of visual archives and argue for a "visual digital turn" in digital humanities. Yet state-of-the-art models require dedicated GPU clusters and cloud computing, making scaling economically prohibitive and environmentally unsustainable. Strubell et al. (2019) estimate that a single neural architecture search produces over 284 metric tons of CO₂ equivalent, five times the lifetime emissions of an average automobile. These findings reinforce the "Digital Humanities and the Climate Crisis" manifesto's claim that "the digital is material" and its call for energy-efficient practices. For researchers seeking sustainable alternatives for computational film analysis, however, few options exist. Pustu-Iren et al. (2020) confirm, "software solutions for the scholarly study of film that utilise video analysis algorithms are still relatively rare," reflecting assumptions about resource allocation. The most advanced surveyed tool, the Distant Viewing Toolkit (Arnold & Tilton, 2020), provides automated semantic extraction via TensorFlow-based CNNs, including face recognition and character-level analysis. Still, CNN inference across thousands of frames remains computationally intensive on a CPU. The field requires not different tools but an entirely new architecture. The `tidylens` package offers that architecture. It uses only classical signal-processing algorithms without deep learning or pre-trained models. Built on minimal computing principles (Sayers, 2016; Gil & Ortega, 2016; Risam & Gil, 2022) — organized around "what do we need?" and Sayers's goal of reducing processing demands and environmental impact — it returns color, shot, and audio extractions as tidy data frames per Wickham's (2014) one-observation-per-row standard. Film scholars can combine their outputs with other packages, such as `tidytext`, `sf`, or `tidymodels`, without additional data wrangling. While these classical algorithms sacrifice CNN-level semantic precision, they reliably measure chromatic and temporal properties and complete full-corpus runs in minutes. Every step is inspectable, consistent with Flueckiger and Halter's (2020) insistence on theoretically grounded interpretation, making classical methods a commitment to transparency, not convenience. To test this approach, the paper applies `tidylens` to a fifteen-film corpus — eight CC-licensed Blender Open Movies and seven public-domain silent films, accessible without institutional licensing. The pipeline ran in under two hours. Measuring colourfulness (Hasler & Süsstrunk, 2003) yields three chromatic categories: six silent films score near zero, as expected for grayscale transfers; Blender films register a median of 29.7; and a print of *The General* (1926) records 12.1. That the corpus sorted into these registers through measurement alone — without annotation or inference — demonstrates classical methods can track authentic material variation. The results show that addressing access inequity and unsustainable computing together is most effective. Classical techniques without specialized hardware reduce energy consumption while enabling the corpus-level analysis that computational film history requires.

This paper presents a screenplay corpus of 2,154 American films (1920-2025), containing 5.2 million classified text blocks organized into over 283,000 scenes, built for semantic search where the scene functions as the primary unit of meaning. The corpus originates from the MovieSum dataset (Saxena & Keller, 2022), extended through a custom PDF-to-JSON parser using spatial analysis of screenplay page geometry. Each block carries a tag type (dialogue, action, parenthetical, scene heading, character, transition), a normalized speaker attribution, and a quality classification. The proof-of-concept corpus spans 2,154 films validated through infrastructure integrity tests, user scenario simulations, and cross-validation between aggregated statistics and raw block-level data. Pre-embedding structural verification revealed that 17.4% of dialogue blocks required lateral join attribution due to intervening parenthetical blocks, a corruption rate that would propagate silently through any embedding pipeline operating on unvalidated input. This segmentation methodology treats the scene boundary as the fundamental articulation point. In screenplay format, the slug line (for example: INT. COFFEE SHOP - DAY) functions as a boundary where dramatic context resets with a new spatial environment, new configuration of characters, and new temporal setting. Embedding at the scene level preserves these articulations. A query for "nostalgia" returns scenes where nostalgia operates dramatically, in dialogue, in described action, in the spatial and temporal framing of the scene heading, rather than isolated sentence fragments that happen to contain semantically adjacent vocabulary. The practical consequence is that the corpus enables queries that produce genuine discovery. A search for scenes where a protagonist confronts institutional corruption with personal moral authority does not return a ranked list of keyword matches. It returns the structural architecture shared across films that have never been studied together: the spatial staging, the dialogue rhythm, the configuration of power between speakers. Connections between films that criticism treats as unrelated become visible because the search operates on dramatic function rather than surface vocabulary. The tool does not replace close reading. It generates the comparisons that close reading can then investigate. This paper positions itself within the lineage of computational film analysis running from Salt's statistical style analysis (1974) through Tsivian's Cinemetrics platform (2005) and Manovich's cultural analytics framework (2020). Where Arnold and Tilton (2023) established a rigorous framework for computational analysis of visual corpora in film and media, this project addresses a complementary gap: the application of semantic computational methods to the textual and dramatic structure of screenplays. The corpus is designed to serve film students, producers, and researchers who need to query screenplay structure semantically, where the integrity of the search unit determines whether the results are analytically useful or noise.

This presentation examines the emergence of AI-powered optical character recognition (OCR) as a transformative solution for unlocking access to historical archives, while critically engaging with the environmental implications of computational choices in digital humanities. Through a case study of 13,848 pages of intercepted French diplomatic telegrams from World War II held by Library and Archives Canada, we demonstrate how emerging AI technologies enable knowledge production from archival materials that would otherwise remain inaccessible, while raising important questions about the environmental costs of such computational approaches. Historical archives face a fundamental accessibility problem: without searchable, structured metadata, researchers cannot effectively discover or analyze their contents. More critically, the combination of technical limitations and resource constraints means that many archival collections are simply never transcribed. Manual transcription remains prohibitively expensive and slow at scale, creating a barrier to knowledge production. The result is that vast quantities of historical materials remain effectively locked away, their research potential unrealized. For the Vichy diplomatic telegram corpus, this meant that despite digitization efforts, the 9,000+ intercepted communications documenting French wartime diplomacy, intelligence operations, and international networks remained largely inaccessible to researchers beyond those willing to manually navigate microfilm reels page by page. Recent advances in multimodal AI models, specifically vision-language models like Mistral OCR, offer dramatically improved accuracy for historical document processing. Our automated extraction pipeline successfully processed the entire corpus, extracting metadata including sender/recipient locations, dates with substantially higher accuracy than traditional approaches. However, this computational approach raises critical questions about environmental responsibility in digital humanities. GPU-based inference for the models used carries measurable energy costs. We must openly acknowledge and critically examine the trade-off between computational energy expenditure and the research value of making archives accessible. When does the environmental cost of AI processing justify the knowledge production it enables? How do we balance the urgency of preserving and opening access to deteriorating archival materials against the ramifications of computing for environmental crises? Our response emphasizes open science principles as partial mitigation of environmental costs. By publishing the complete processed dataset under a Creative Commons Attribution 4.0 license via an Omeka digital collection platform (https://omeka.uottawa.ca/examination-unit/), implementing IIIF standards for image interoperability, and archiving data in open repositories with DOI assignment, we ensure that other researchers can reuse our outputs without requiring redundant processing. This approach reduces duplicated computational expenditure while maximizing the research value extracted from the initial energy investment. Our methodology documentation is published openly, enabling critical evaluation and adaptation by other projects facing similar archival challenges. This project contributes to urgent conversations in digital humanities about sustainable computing practices. We argue that the field needs frameworks for measuring, documenting, and justifying the environmental costs of computational choices. Rather than avoiding AI technologies entirely or adopting them uncritically, we advocate for transparent assessment of trade-offs between accuracy gains, knowledge production potential, and environmental responsibility. The emergence of powerful AI tools demands equally robust emergence of ethical frameworks for their deployment in cultural heritage contexts.

This paper proposes a Critical Digital Humanities intervention into the algorithmic surveillance of Indigenous geographies. I stage a hauntological dialogue between the "blind" InSAR satellite—which filters atmospheric moisture as "noise"—and Rukai lands in Taiwan, where that very moisture (wakokongo) constitutes a sacred medium of temporal knowledge. Drawing on Joseph Beuys's 1965 performance, where honey bridged rational language and animal flesh as a "conductive material," I reframe discarded atmospheric data as "Digital Honey": a materially resistive medium capable of bridging technocratic vision and Indigenous cosmology. Methodologically, I develop speculative sonification through three operations: 1. Forensic Excavation: Using ESA's SNAP toolbox, I aim to reverse-engineer atmospheric correction algorithms to recover deleted phase delay data—the "digital waste" defined as noise. 2. Sonic Translation: I convert this data into acoustic waveforms, creating what I term acoustic hauntologies. This reveals monsoon rhythms—precisely what Rukai cosmology attends to as wakokongo. 3. Cosmopolitical Reassemblage: I reimagine risk maps as cosmopolitical assemblages integrating atmospheric affects. This act of "neuroqueering the dataset" challenges violent scalar dislocation: the achievement of "planetary" coherence through erasure of Indigenous "ground" temporalities. This project contributes to decolonial digital humanities by demonstrating how aesthetic coding practises can recover silenced ontologies from within planetary governance infrastructure. By making algorithmic "waste" audible, I propose sonification as epistemic reparation.

As generative AI systems rapidly reshape knowledge production, Digital Humanities (DH) confronts a dual condition of emergence and emergency. On one hand, large language models (LLMs) promise new forms of multilingual research, cultural analysis, and collaborative scholarship; on the other, these systems operate within global data economies marked by structural asymmetries in language representation, computational power, and infrastructural control. This paper argues that the rapid application of LLMs in cultural preservation initiatives constitutes an epistemic emergency in cultural studies, particularly in relation to minority and Indigenous languages across and beyond the Americas. Drawing on the framework of data colonialism and engaging scholarship in critical data studies, postcolonial theory, and Indigenous data sovereignty, the paper examines how English-centric LLM systems, together with respective private and public stakeholders appropriate, reorganize cultural and linguistic data within extractive technological regimes. Case discussions involving Māori, Guaraní, Inuktitut, and Irish language initiatives illustrate recurring tensions between technological innovation and community governance. While large language models are often presented as tools for preservation and revitalization, they may simultaneously centralize authority in corporate infrastructures and reproduce hierarchies embedded in training data. Rather than rejecting outright the use of LLM for cultural preservation purposes, the paper proposes a governance-oriented model for DH and its extended projects grounded in corpus accountability, LLM specificity, infrastructural transparency, and community participation in model design and evaluation. By foregrounding the CARE principles and debates on data sovereignty, it reframes LLMs as a site of negotiation over epistemic authority. Positioning DH within broader transnational conversations about equity and justice, the paper proposes that responding to AI applications across different scenarios or disciplines requires solidarity across linguistic and geopolitical contexts. The challenge is not only technical but ethical and institutional: to determine how emerging computational systems can support, rather than subsume, diverse knowledge traditions.

This paper argues that meme scholarship emerging from the Global South offers not only new objects of study for digital humanities, but an emergent methodology shaped by conditions of precarity, multilingualism, and platform asymmetry. Responding to ACH 2026’s theme of Emergence/ia, I examine Indian meme cultures on Instagram as sites where creativity unfolds within ongoing economic and infrastructural emergencies. Rather than approaching memes primarily through scale, virality, or large-scale data extraction—frameworks that often dominate Anglophone digital humanities—this paper repositions memes as processual, performative practices through which users collectively negotiate crisis within platformed environments. Methodologically, the study integrates three components. First, close visual and textual analysis of meme templates, captions, and comment threads examines multimodal construction, irony, and repetition. Second, discourse analysis informed by Judith Butler’s theory of performativity conceptualizes memes as reiterative acts that produce and stabilize social identities—particularly middle-class aspiration—through stylized repetition. Third, a platform-aware analysis investigates how Instagram’s affordances—algorithmic feeds, shareability, follower metrics, and remixable templates—condition visibility and shape the circulation of humor. Rather than privileging purely quantitative scraping, the project employs situated sampling and reflexive digital ethnography to foreground relational engagement and community-specific humor practices. The findings reveal three interrelated dynamics. First, “middle-class aspiration”memes function as performative reiterations of economic anxiety: humor and exaggeration normalize precarity while subtly critiquing neoliberal promise. Second, Hindi-English code-switching within meme captions and comments fosters affective solidarity and challenges assumptions of monolingual digital publics, offering a model of multilingual digital emergence. Third, Instagram’s algorithmic visibility structures which narratives of aspiration circulate widely, demonstrating how creative emergence operates within infrastructural constraints. Together, these findings suggest that Instagram meme cultures in the Global South operate as vernacular humanities infrastructures—collaborative spaces where users theorize class, belonging, and survival under conditions of emergency. By foregrounding process over product and relational analysis over scale, this paper proposes meme scholarship as an emergent digital humanities method capable of decentering U.S.-centric epistemologies and rethinking knowledge creation across transnational contexts.

At a deeply unsettling moment globally and in the U.S., and with recent changes to key funding institutions in the humanities, emergency planning has become more crucial than ever. At the University of Nebraska-Lincoln's Center for Digital Research in the Humanities (CDRH), this planning has included the sometimes unglamorous-feeling work of streamlining processes, updating hardware, and simplifying websites and site frameworks. Founded in 2005, the CDRH has helped to build and host dozens of digital humanities websites. The CDRH's leadership, staff, affiliated faculty, and position within the organizational structure of the UNL Libraries have all changed over the course of this time, and many different tools have been adopted and deployed as new technologies and standards emerged, new programmers brought different expertise, and new projects were developed. In the past few years, the Center has launched a concentrated effort to revisit core software and reduce the number of frameworks in active use for digital humanities projects. Like minimalization generally, and in keeping with preservation guidelines like those issued by the Endings Project, this work is intended to help with long-term maintenance and sustainability. Changes were balanced and combined with ongoing work to develop projects as well as a related initiative to improve accessibility for all Center websites. In this talk we will discuss some of the particulars of this effort at the CDRH, including some of the legacies of a twenty-year-old digital humanities center, challenges related to libraries and file formats, unresolved questions, and opportunities presented by new tools and platforms.

While emerging conversations around accessibility are far from uncommon in information and library science, accessibility scholarship often neglects to clearly define it. The growing shift in digital archival praxis demands we question what accessibility means in these spaces. In particular, the updated ADA Title II makes exemptions for “archived” content, providing the content is kept “only for reference, research, or recordkeeping… in a special area for archived content… [and] has not been changed since it was archived” (ada.gov, 2024). While archives should and do have a responsibility to make their content accessible, the language of Title II provides an exemption for many. In hopes of simplifying this process and encouraging digital archives to make collections accessible, we propose the ENDURS framework (embodiment, navigability, discoverability, usability, readability, sustainability), to improve archival accessibility at all levels. “Accessibility” often is used interchangeably with “visibility” and “usability”, which muddies the water on how to effectively and ethically implement accessibility protocols in LIS. For example, Mukwevho and Ngulube (2022) address visibility in public South African archives using “accessibility” interchangeably with discoverability and visibility. Friedewald et al. (2024) examine policy recommendations on proving access in public archives, similarly using “accessibility” to discuss enhancing user access. To ensure clarity in future scholarship, we distinctly define each of these words and showcase why they are important in the field of archives. Using aspects of Universal Design (Spina, 2022)–the theory that archives should be accessible to all people, Complex Embodiment Theory (Baucom, 2024)–the belief that our bodies and experiences shape our identities, and cultural competence (Michalak & Ellixson, 2025)–the study of communication across peoples and groups, we apply the ENDURS to three domestic and three international digital archival projects. Utilizing this framework, we examine both global and domestic digital archival projects, identifying weaknesses in Title II recommendations and traditional archival practice, to recommend actionable accessibility guidelines and features. Our findings establish foundational principles that will help bridge the gap between digital archives in varying cultural, institutional, and linguistic contexts.

In the months leading up to April 2026, campus-wide emails remind faculty that all teaching materials must meet federally legislated standards for digital accessibility. Official communications re-state the requirement with a list of links, while technical explanations are often left to instructional designers with little personal experience with interfacing with a diverse population of students, or campus IT professionals who assume all teaching materials are simply “content” to be “delivered”—and only ever in English. All are doing what they can, but when the responsibility for making digital materials accessible is essentially (dis)placed on to individual faculty, the institution is abnegating its responsibility for systemic issues and avoiding opportunities for centering access. What, then, can digital humanities pedagogy contribute in a responsive and collaborative way? Using my public research university campus as a case study, I reflect on the higher-ed pedagogical situation and suggest possible moves, hopefully opening conversation across institutions about this issue that many have been facing with a sense of professional isolation, especially precariously-employed adjuncts who fear being held individually accountable. (Recall Robert McRuer’s arguments that, on a global scale, disability is central to austerity politics—and to the rhetoric of emergency used to justify them.) Those who are most attentive to languages and images, the humanities faculty teaching languages, art history, film, and visual studies, are the most concerned with the digital distribution of the multilingual media they teach. And those in DH-informed pedagogy are well positioned to help cultivate the much-needed relationships of practical knowledge-sharing. The institutional emphasis on legal compliance in what is presented as an exigent situation (with a calendar deadline and vague forebodings of enforcement) is thus also emerging as an opportunity for the cultivation of what Margaret Price calls “collective accountability and gathering.” How can habits of thinking as well as working be shifted in thoughtful ways, using the insights of critical disability studies, multilingual DH, and pedagogies of care, toward inclusive practices? Jay Dolmage, who has long worked to help teachers implement Universal Design not through checklists but through critical engagement, has pointed out the limitations of university administrations’ paying lip-service to Universal Design in teaching without investment in the campus resources that could actually remove barriers for students with disabilities. Moving beyond the checklist or technical quick fix, then, how can we cultivate what Aimi Hamraie calls “access-knowledge” on our own campus? While avoiding what Ashley Shew calls “technoableism,” how can we take the opportunities for exploring technology alternatives, for example for OCR or speech-to-text, rather than re-entrenching the unquestioned use of barely functional commercial software and platforms? Opening these generative questions, I will share examples of small-scale praxis (sometimes successful, sometimes not), including mini-workshops, short video demonstrations, and the creation of minimal resource documents over which colleagues can “gather.” This presentation will include reflection on experiences in the spring of 2026 that I can’t fully predict, but for which scholarship in disability studies and education provides a strong and flexible framework.

This paper uses computational text analysis to examine whether a “Crazy Rich Asians (CRA) effect” can be detected across Malaysian novels published in the period following the novel’s release in 2013 and the 2018 film adaptation. This “CRA effect” alludes to the phenomenon where, increasingly, novels borrow from the themes and genre conventions made popular in association with Southeast Asian novels by CRA. Such an effect is most apparent in novels like The Last Tang Standing, which is described in its marketing paratext as “Crazy Rich Asians meets Bridget Jones's Diary”, or the Indonesian novel Dial A for Aunties, which is described by one reviewer as “a mix of a J Lo rom-com, Kevin Kwan’s Crazy Rich Asians opulence, and Weekend at Bernie’s” (Simone and Her Books). The paper asks: to what extent can computational methods detect CRA as a force of literary commensuration for novels published in English by Southeast Asian authors, particularly Malaysian authors? The study draws from an original dataset of more than 100 novels published in English by Malaysian authors, spanning approximately seventy years, including metadata and the texts themselves. A preliminary examination of the metadata suggests that most of the novels published by US and UK presses in the dataset appear after 2013. Prior to 2013, historical fiction focused on war, memory, and colonial violence appears more frequently. After 2013, a larger number of contemporary novels, often oriented toward romance and urban life, enter global circulation. While this temporal clustering does not on its own establish causation, it prompts further inquiry into what kinds of stories become more marketable and more tied to the genre of “Southeast Asia” in the Anglophone publishing world, in the wake of CRA’s global success. To investigate whether such an alignment is detectable at the level of textual features, I compile and digitize more than 100 Malaysian novels and deploy a set of text mining methods that allow me to compare novels published before and after 2013. These methods include distinctive-word analysis and classification-based similarity measures, using the CRA novel as a reference text, enabling me to analyze whether CRA has changed the center of gravity for Anglophone Southeast-Asian novels in the global market. Building on work in computational literary studies by scholars such as Ted Underwood and Hoyt Long, as well as sociological analyses of literary markets such as Graham Huggan’s The Postcolonial Exotic, the study demonstrates how DH methods can operationalize and measure processes of literary commensuration across time. More broadly, the paper contributes to digital humanities by offering a methodological framework for using computational techniques to detect market-driven shifts and pressures in genre formation and thematic convergence, responding to longstanding inequities in the global literary market for literatures from the Global South.

Duke University’s AsteXT project uses Named Entity Recognition (NER) and BERT fine-tuning to parse Asian American short stories (1974-2024). The AsteXT model traces how “space” has been portrayed across these five decades of literature, the authors of which bring their own literary expressions from a range of Asian and Asian American experiences. Using the team’s manual fine-tuning process as an example, this paper argues that as language models rapidly advance (and thus become more abstract), Computational Literary Studies (CLS) scholars must operate more intimately with their corpus to develop a domain-specific understanding strong enough to counterbalance the abstractive tendencies of a model. While data annotation is seen as too resource- or time-intensive, prone to inaccuracy, and difficult to perform at scale, we argue that a systematic approach to manual data annotation is crucial for more responsibly and accurately parsing minority literature. Given that such literature often follows a linguistic and historical schema unique to its minority creators, our corpus has demanded an approach that moves beyond the schemas produced by dominant cultures or communities and used to train our industry standard LLMs (see Naous, et. al.’s 2024 paper “Having Beer after Prayer?”). Our argument therefore reenvisions the utility of data annotation insofar as we see it as vital–despite its demands and risks–to the quality of any model processing non-dominant literary forms. When manual annotation follows a systematic, iterative, and replicable workflow that centers the cultural and historical contexts within our corpus, it allows for future researchers to work more confidently with a minority literary dataset. Building on previous NER work in Evans and Wilken’s 2018 paper “Nation, Ethnicity, and the Geography…”, we present a new NER schema: Soft NER spatial recognition. “Soft NER” focuses on collecting non-Named Entities, like improper nouns, but with culture-specific connotations. Our Soft NER framework contains five categories, which allow us to meet the range of spatial expressions in our corpus: Private space, Communal space, Extraterrestrial/Figurative space, Natural space, and Institutional space. For example, traditional NER tools like spaCy do not identify the term “heavens,” a plural improper noun, in any spatial category, though with a strong understanding of the religious and cultural connotation of “heavens” in Asia, we tuned our model to recognize “heavens” as an Extraterrestrial/Figurative space. Ultimately, this paper argues that a systematic and manual approach to textual processing can, among other benefits: meet urgent needs for concrete, community-specific criteria for processing creative works; develop a stronger sense of care and stewardship of the data; cultivate an intellectual respect for the unique historical and cultural contexts from which minority-created data emerged over time; and react flexibly to data by transforming existing Natural Language Processing tools in ways that reflect an intellectual care and respect. By sharing our unique methodology based on So and Roland’s reflexive philosophy in “Race and Distant Reading” (2020), we provide CLS scholars not just a new NER schema, but also a model for introducing humanistic methods that close the gap between texts and digital tools.

Minimal computing emerged within digital humanities as both a technical and political stance: a commitment to building projects that are low-cost, low-bandwidth, accessible, and sustainable, particularly in contexts marked by infrastructural inequality (GO::DH Minimal Computing Working Group, 2017; Gil & Risam, 2019). Rather than privileging scale, novelty, and computational abundance, minimal computing prioritizes sufficiency, maintenance, and care. It resists techno-solutionism by asking not what is technically possible, but what is necessary, just, and sustainable within specific institutional and environmental conditions. In the age of large-scale artificial intelligence, this orientation becomes urgent. Contemporary AI development, such as large language models (LLMs), has been defined by escalating computational demands, rising carbon footprints, and increasing concentration of power (Crawford, 2021; Brevini, 2022). Given the proliferation, or hype (Bender, 2025), of AI systems, can the principles of minimal computing be meaningfully applied to AI projects? And if so, what would that require? To explore this question, we bring minimal computing into conversation with three emerging strands of AI discourse. First, Green AI argues that efficiency and computational cost should be treated as primary evaluation criteria alongside performance metrics (Schwartz et al., 2020). Second, Small Data AI proposes moving away from large-scale data dependency toward techniques capable of operating under scarcity (Orrù, 2025). Third, traditions of AI refusals articulated in feminist and justice-oriented critiques of extractive data regimes (Feminist Data Manifest-No, 2020) which insist on the legitimacy of saying no to systems that reproduce surveillance and inequity. We then propose principles that are particularly important for Libraries, Archives, and Museums (or LAM) institutions. In contrast to corporate AI’s extractive paradigm, LAM institutions can model slower, care-centered AI practices that prioritize interpretability, sustainability, and community governance. Minimal computing, we argue, does not reject AI outright but rather frames it as a question of sufficiency, responsibility, and power.

Drawing on the 2015 chapter by Ian Gregory, David Cooper, Andrew Hardie, and Paul Rayson, “Spatializing and Analyzing Digital Texts: Corpora, GIS and Places”, we developed a two-stage activity for participants in the Mellon-Funded Digital Borderlands in the Classroom initiative. In this presentation, we will discuss how our activity introduces principles from applied linguistics and geospatial analysis to our participants. In the first stage of the workshop activity, the digital scholarship specialist lays the groundwork by introducing the lightweight text analysis platform Voyant Tools, introducing distant/close reading practices on a curated collection of historical newspapers from the US-Mexico border. As part of the exploratory corpus analysis activity, participants were asked to identify three locations within the newspapers that contained some form of geographic identifier(s) that could be used for the Geographic Information Systems (GIS) portion of the workshop. The second stage of the workshop, led by the geospatial specialist, began with introducing participants to the concepts of geocoding using location information to create a spreadsheet with descriptive as well as spatial information. Participants then used Esri’s ArcGIS Online to plot the locations on an online map to create a geospatial visualization of the corpus analysis that combined descriptive and spatial attributes. This multi-stage workshop, combining linguistic and geospatial data collection, analysis, and visualization techniques, introduced faculty fellows to a number of methodological concepts (e.g., distance/close reading, corpus analysis, geocoding, mapping) as well as number of software programs (e.g., Voyant Tools and Esri ArcGIS Online). Works cited Gregory, Ian, David Cooper, Andrew Hardie, and Paul Rayson. 2015. “Spatializing and Analyzing Digital Texts: Corpora, GIS and Places.” In Deep Maps and Spatial Narratives, edited by David Bodenhamer, John Corrigan, and Trevor M. Harris. Spatial Humanities. Indiana University Press.

Over the last two decades, the United States has increasingly installed authoritarian mechanisms into domestic governance, policing, and social institutions. Today we see masked ICE agents whisk undocumented immigrants and political opponents into murky detention facilities while, behind the scenes, private military contractors monitor social media for expressions of dissent. Meanwhile, the National Guard and unspecified “federal agents” patrol major cities under the guise of stifling crime, and the state tightens control over news media by denying reporters access to press briefings, revoking the licenses of major news agencies, and arresting reporters covering public protest. The world watches with varying degrees of alarm, but each of these tactics has simmered out of sight of the mainstream for decades. Building on the work of scholars such as Ruth Wilson Gilmore, Nikhil Pal Singh, Stuart Schrader, and Radley Balko, this paper examines how racialized policing combined with emerging markets in private military contracting after 9/11 to create elaborate webs of public-private partnerships capable of skating legal safeguards for repressive purposes. The alliances between police, para/militaries, government, corporations, lobbyists, and private citizens power the ongoing emergencies across the nation. However, these collaborations emerged—and proved efficacious—through deployments against the #NoDAPL and #BlackLivesMatter movements of the 2010s. While situated within the ongoing conversation of race, policing, and militarization, this paper also utilizes social network analysis and sousveillance, investigative journalism, Indigenous scholarship, and a cache of internal private military documents to unmask the infrastructure of repression tested a decade ago and unleashed with abandon today. This talk delves into the collaborations leveled against the #NoDAPL anti-pipeline movement between 2015 and 2017 in an effort to map the wide range of political, social, economic, and legal institutions bound together in pursuit of quashing dissent. The #NoDAPL case study offers unique insight into the current American landscape by exposing the (now) recycled blueprint of surveillance, legal campaigns, and shadowy partnerships that buttress ICE’s unfolding reign of terror. By stepping into the past and mapping the social network of repression, this presentation ultimately provides inroads for understanding and dismantling key tactics of information operations, militarized repression, and weaponized legislation utilized against social justice activists and opponents of America’s authoritarian turn.

Mapping the Black Digital and Public Humanities (Mapping BDPH) is a searchable database of over 700 Black Studies projects in the digital and public humanities, including multiple dozens of projects on Black and multiethnic literatures. Emerging out of my experience working on small local Black historical and literary recovery projects, Mapping BDPH was imagined with two mutually reinforcing goals in mind. First, finding: to make it easier to find digital and public projects about Black history and culture, amplifying the hard work of fellow project creators. And second, founding: to make it easier to see gaps, intersections, and possibilities and thereby found coalitions across similar projects—all with the aim of fostering deeper learning from one another’s strategies and building more meaningful connections. This paper considers Mapping BDPH through the lens of emergencia, taking seriously the dual meanings of emergence as growth, connection, and creation, and emergency as crisis, urgency, and response. Drawing on scholarship in Black Digital Humanities, decolonial digital public history, and community‑engaged infrastructure, I examine how the project has evolved in response to current political conditions—particularly through our current and ongoing efforts to better include Afro‑Latinx digital and public humanities work. DH projects such as First Blacks in the Americas, the Dominican Voices Project, Taller Entre Aguas, and the Black Central Americas Project exemplify forms of hemispheric Black knowledge production that challenge U.S.‑centric frameworks. But incorporating these projects into our database, and further expanding to include non-English projects based outside of the U.S., raises practical and ethical questions central to ACH 2026’s focus: How do we meaningfully connect with—and connect—multilingual projects from around the world? How do we address the ethical issues connected with amplifying the work of these practitioners without unintentionally exposing them to hostile audiences or harm? And how do we develop productive alliances and funding strategies to support this work in times of increasing political and economic crisis and uncertainty? Rather than offering simple solutions to these questions, I consider Mapping BDPH as a case study in relational DH practice—one that grows through dialogue, iteration, and community input. I also explore why, in this time of emergency, Mapping the Black Digital and Public Humanities—and the grassroots, emergent ethos of finding and founding that informs it—are now more important than ever.

What is the relationship between the digital humanities and the technology industry? In particular, how can digital humanities practitioners cultivate a relationship with the largest technology companies and their products that reflects our own goals and commitments? This presentation proposes as a model an orientation explored by Agnès Varda in the 2000 film Les Glaneurs et La Glaneuse (The Gleaners and I). Varda’s documentary both explores and embodies the practice of gleaning. The gleaner follows the harvest, gathering and collecting food that is left behind. Varda films gleaners at work, but as the film’s French title makes clear, Varda considers herself a gleaner as well, assembling her film from what others have left behind. When applied to digital humanities practice, Varda’s enriched conception of gleaning has great potential as an orientation towards the technology industry and its products. The contemporary technology industry can be characterized by its extraordinary level of investment in companies and products during the early stages of their commercial application, and the resulting proliferation of free and subsidized services and products. These can include consumer-facing products like chatbots and social media sites, as well as open source tools and frameworks – for example, web development foundations like Bootstrap, React, and Angular – developed within the technology industry but widely reused in both academic and commercial settings. Whether in research, pedagogy, or everyday life, these free, advertising-supported, and subsidized products are nearly unavoidable. What are the implications of incorporating these corporate left-behinds into the infrastructure of digital humanities practice? Just as Varda styles herself and other artists featured in her film as gleaners, creating meaning and beauty from abandoned objects and stories, perhaps the concept of gleaning can complement established approaches like minimal computing and inspire new approaches to the design of minimal infrastructures. Varda’s conception of gleaning has informed the development of Rondo, a minimal framework for building digital exhibits and projects, which allows users to build interactive sites using a Google Sheet or Excel spreadsheet. Rondo builds on minimal computing frameworks like Wax and Collection Builder, but it uses a gleaning-inspired single page application architecture which allows the spreadsheet to serve as a real-time back-end database for the site. Rondo was developed at San José State University in the King Library Digital Humanities Center with the goal of supporting classroom and community digital humanities projects. The first production project to use Rondo - Running in Prayer: California 500 Mile American Indian Spiritual Marathon Relay Team - was released in Fall 2025. Rondo is an example of the creative possibilities the gleaning approach presents. The way we conceive of our relationship to technology can inform the possibilities we imagine, whether those possibilities take the shape of infrastructure and tools or scholarship and projects. This presentation proposes that Varda’s conception of gleaning can be a generative and fertile way for digital humanities practitioners to position themselves in relation to the technology industry, and to engage with technology in ways that preserve our autonomy and serve our highest goals as researchers and individuals.

Released in January 2016, the Epstein Files offer rich context of Epstein’s network in politics, business, entertainment and science. While systematic sexual abuse and exploitation of women and girls are evident in this case, the extent of Epstein's network is poorly understood. Journalistic reports and digital projects have uncovered some of the stories, but to exhaust a volume of 3.5 million pages, 2,000 videos, and 180,000 images requires more than close reading or keyword search. Digital humanities can thus facilitate document processing and information extraction in many ways. This research aims to build a visualized network of the Epstein Files based on the emails from and to Jeffrey Epstein in the disclosed text files, showing the persons involved, the nature of relationships in terms of email exchanges, and the topics in email contents. Digital humanities methods such as OCR, NER and topic modeling will be applied in this research. Data visualization tools such as nodegoat will also be explored. Method criticism will be conducted along the way to analyze tool efficiency and evaluate the ways in which contemporary legal documents are processed. When submitting this proposal, the author has done a primitive NER analysis of 181 email files in the Epstein Files, identifying persons in the files. The most common name that surfaced is “Boris Nikolic”, previously a science advisor to Bill Gates who was also listed as an executor of Epstein’s will. Although Boris denied knowledge of the will and said he was shocked (Fox Business, 2026), a close look at the emails found he was in close contact with Epstein from 2010 to 2016, and was a key figure in the Epstein network. (Wall Street Journal, 2026) By expanding the scale of texts and critically applying DH methods, the author expects more revealing findings in this research. References: U.S. Department of Justice (2026) Epstein case materials. Available at: https://www.justice.gov/epstein (Accessed: 22 February 2026). Fox Business (2026) Who is Boris Nikolic, Epstein’s executor? Available at: https://www.foxbusiness.com/features/who-is-boris-nikolic-epstein-executor (Accessed: 22 February 2026). The Wall Street Journal (2026) How Epstein inserted himself in a split between Bill Gates and a top Gates adviser. Available at: https://www.wsj.com/articles/how-epstein-inserted-himself-in-a-split-between-bill-gates-and-a-top-gates-adviser-6d8fca69 (Accessed: 22 February 2026). Roll Call / FactBase (2026) Epstein file: House Oversight 026764–026766. Available at: https://rollcall.com/factbase/epstein/file?id=House_Oversight_026764-026766 (Accessed: 22 February 2026). Google Journalist Studio – Pinpoint searchable database. (n.d.) Available at: https://journaliststudio.google.com/pinpoint (Accessed: 23 February 2026). Epstein Archive (n.d.) Epstein Archive – Document browser. Available at: https://epstein-docs.github.io/ (Accessed: 23 February 2026). Andrews, M. (n.d.) Epstein Doc Explorer. GitHub repository. Available at: https://github.com/maxandrews/Epstein-doc-explorer (Accessed: 23 February 2026). Paul, G. (n.d.) Epstein Document Search. GitHub repository. Available at: https://github.com/paulgp/epstein-document-search m (Accessed: 23 February 2026). Steininger, E. (n.d.) Epstein Docs. GitHub repository. Available at: https://github.com/esteininger/epstein-docs (Accessed: 23 February 2026).

This research paper states, the complex digital mourning phenomenon and how social media profiles become, what can be termed 'emergent archival systems' after the death of a user, is clearly the focus of this paper. This is related to the conference theme of Emergence/ia. In this examination of digital remnants, the study identifies that the duality of digital remnants provides emergent sites for new forms of collective grief and memory development; while simultaneously representing emergent sites for the fragility of digital systems, and therefore, the durability of digital memorials. The study draws on theory from digital cultural heritage and thanatosociology to investigate a body of memorialized profiles, or 'digital shrines', on both Instagram and Facebook; by using a hybrid approach comprised of computational text analysis to identify emotionally charged language and instances of grief, which are paired with a close reading to describe the affordances and constraints presented by platform governance. Through this analysis, the study is able to illustrate that 'digital objects', particularly those represented by the final posts, stories or comment threads of the deceased, necessitate the process of sacralization. Further, the study suggests that these digital objects should not be reviewed as fixed artifacts, but rather as dynamic objects requiring an entirely new review of curatorial labour. The study brings attention to the often hidden emotional labour necessary for the preservation of digital legacies as a result of cloud storage or platform failure. In this paper, the study is able to disclose why it is an "emergency" that there will be no digital preservation. The idea that platforms exist as uncurated containers of human memory provides the potential for the destruction of that memory and therefore creates an ethical crisis for digital humanities scholars. The study looks at how communities are responding to this crisis of urgency by "digital solidarity" to rescue the memories lost within the algorithmic flux, including using third-party archival tools and developing cultural practices around taking screenshots to capture the memories. Additionally, this paper provides a contribution to the study of humanistic research about digital cultures through this new model of "critical care" for what is being called digital afterlives. Rather than seeing an archive as having a fixed location in space (e.g., a library), the study argues for the need to view the archive within a context such as the changes in the use of archive platforms over time, especially given that many of these new emerging practices of memorializing/remembering/dealing with loss in digital spaces/services have evolved in the context of social media. Through this analysis, the study demonstrates how digital humanists can respond to the realities of loss in the digital age and help ensure that these emergent practices of mourning on the web are duly acknowledged as part of our cultural heritage.

Bimodal network graphs create rich distant readings of complex literary-historical relationships, as Liz Fischer has demonstrated in Network Analysis for Book Historians. Thoughtfully identifying two kinds of entities in relationship to each other — manuscript and text, text and editor, book and owner — enables multiple kinds of graph projection and network analysis. In this paper, I present two bimodal networks of persons and books in late eighteenth-century British publishing. Specifically, I examine the hundreds of “subscribers” whose names were recorded as financial supporters of publications by three Black Britons. Ignatius Sancho, Olaudah Equiano, and Ottobah Cugoano all essentially “crowdfunded” their works in a process known as publishing by subscription — so called because the patrons subscribe their name to a list, offering tantalizing evidence for a book history of ordinary readers. I have developed a Python codebase using highly tailored, rules-based systems to build network graphs from transcribed lists. My bespoke system for extensive normalization of names enables, for the first time, thorough identification of subscribers who supported multiple books. My first network contains the thirteen editions published by Sancho, Equiano, and Cugoano, revealing surprisingly little overlap in their audiences but demonstrating two changes Equiano’s publishing strategy between editions. The second graph, of all digitized subscription publications from 1782 (thirty-four of forty-two), shows how Sancho's book aligns with a specific subset of literary patrons. Ultimately, both networks demonstrate exciting results about these authors, gained through a computational process that can now be scaled to any eighteenth-century subscriber list. Selected Bibliography Ahnert, Ruth, Sebastian E. Ahnert, Catherine Nicole Coleman, and Scott B. Weingart. The Network Turn: Changing Perspectives in the Humanities. Cambridge UP, 2020. Evalyn, Lawrence. "Ignatius Sancho's Subscribers." Cambridge Companion to Ignatius Sancho, edited by Nicole N. Aljoe and Kristina Huang, Cambridge UP, forthcoming 2026. Evalyn, Lawrence, and Alisa Chen. "Book Subscribers of Formerly Enslaved English Authors, 1782–1794." Journal of Slavery and Data Preservation, vol. 6, no. 4, 1 Dec. 2025, pp. 267–77. Fischer, Liz. Network Analysis for Book Historians: Digital Labour and Data Visualization Techniques. ARC Humanities Press, 2025. Hanley, Ryan. Beyond Slavery and Abolition: Black British Writing, c. 1770–1830. Cambridge UP, 2018. Hill, Mark J., et al. "Reconstructing Intellectual Networks: From the ESTC's Bibliographic Metadata to Historical Material." Proceedings of the Digital Humanities in the Nordic Countries 4th Conference, vol. 2364, 2019, pp. 201–19. King, Julia. "Syon Abbey's Books and the Strength of Weak Ties." Manuscript Studies: A Journal of the Schoenberg Institute for Manuscript Studies, vol. 9, no. 1, 2024, pp. 33–71. Pink, E. E., “Frances Burney’s Camilla: ‘To print my grand work … by subscription,’” Eighteenth-Century Studies, vol. 40, no. 1, 2006, pp. 51–68. Rezek, Joseph. "The Print Atlantic: Phillis Wheatley, Ignatius Sancho, and the Cultural Significance of the Book." Early African American Print Culture, edited by Lara Langer Cohen and Jordan Alexander Stein, U of Pennsylvania P, 2012, pp. 19–39.

The rapid advancement of generative AI is a sonic emergency for Black people, as language, mannerisms, and soundscapes are mimicked and recreated with little or no connection to the communities on which they were trained. In this paper, I propose sampling as a technology that can help digital humanists and music scholars begin to counter these harms. As a sonic practice, sampling is best known for its roots in hip-hop music, where DJs and producers have perfected the art of using sections of previously recorded material in new compositions. The ethics of sampling are often limited to conversations about copyright law, intellectual property, and fair use, as artists work to “clear” samples with the estates, record companies, and archives that hold copyright for musical catalogues. However, as deepfakes, memory machines, and other forms generative AI are continually blurring the line between the living and the dead, I offer “sampling with critical intention” as a means of charting a course toward understanding the sonic ethics of a malleable past. In this paper and listening session, I present soundscape compositions from my upcoming project Sampling for Black Life in order to think through how sampling might be an energizing response to rapid technological change. This project samples archival collections from the Library of Congress in order to develop a practice of sampling these materials with rigorous care. Further, drawing on the work of Tonia Sutherland, Jessica Marie Johnson, and Saidiya Hartman, I engage digital audio workstations as tools that can be used within digital humanities to engage larger questions of Black sound, manipulation, fabulation, and repair.

Across Colombia and the Latin American diaspora along the West Coast of the United States, cumbia functions not as a single genre but as a relational field of sound, memory, and belonging. Born in Colombia’s Caribbean region and shaped by Afro-Indigenous and peasant traditions, cumbia has traveled through internal migration, recording industries, radio, and digital platforms to become a translocal language of identity. From Montes de María (one of its emblematic regions) to Bogotá and onward to diasporic communities in California and across the U.S. West Coast, cumbia expresses rural memory, urban transformation, migrant nostalgia, and intergenerational continuity. Its greater power is its adaptability: ritual and popular, acoustic and electrified, archived and remixed, yet consistently recognizable as a shared sonic horizon across borders. In the current socio-political climate (marked by anti-immigrant rhetoric, legal precarity, and renewed threats of displacement) cumbia operates as a crucial medium of self-expression, community-building, and collective resilience for Colombian and Latin American migrants. It sustains affective ties to homeland while enabling new forms of belonging in complex social landscapes. Therefore, preserving and systematizing these materials is not only an archival project but an urgent intervention in safeguarding community memory and expressive life under conditions of vulnerability. This project proposes a “Community-Centered Cumbia Digital Archive” connecting Colombian territories and their diasporic extensions in the western United States. Conceived as a participatory digital database, the archive documents diverse expressions of cumbia while resisting extractive models of cultural preservation. Intergenerational musicians, collectors, DJs, cultural organizers, and community members contribute recordings, photographs, videos, and oral histories, situating songs within lived and migratory memory. The archive thus becomes both research infrastructure and platform for translocal dialogues. The project builds on ethnomusicological and Latin American cultural scholarship that understands music as social practice and diasporic archive (e.g., Peter Wade; Ana María Ochoa Gautier; Deborah Pacini Hernandez; Darío Blanco Arboleda; Héctor Fernández). It also draws methodological inspiration from the Flamenco Letra Database Project led by Tania Flores at Stanford University in collaboration with scholars focused on digital humanities, sound, transatlantic thinking, and archival projects, which demonstrates how community-informed digital translocal infrastructures can preserve vernacular traditions while foregrounding interpretation, lived experience, and community. Similarly, this archive will combine structured metadata (titles, lyrics, recording data, performers, geography, instrumentation, format) with relational narratives that trace how repertories acquire new meanings in migrant settings. Integrating ethnographic fieldwork, digital humanities tools, and participatory archiving practices, the Cumbia Digital Archive positions cumbia as an affective archive of displacement and resilience. It is akcnowledged not only as a genre, but also as a network of belonging where migrant practices and communities sustain life across borders.

This paper presentation will discuss a case study at Appalachian State University that documented the Hurricane Helene natural disaster in 2025 through a website-based crowdfunded community archive project. Natural disasters are increasingly common, and crowdsourced data from citizens can be a vital mechanism for historical documentation. Crowdsourced community archives are easily accessible by the public, but they open the door to significant legal and privacy challenges if not structured thoughtfully. Some of those legal challenges include copyright, privacy, data reuse, third party content, orphan works, sensitive information, defamation, oral history consent, website liability, and perpetual public access, among others. This presentation will discuss the specific legal challenges inherent to the Hurricane Helene Archive case study in further detail and offer solutions to mitigate legal risk and ensure the long-term use of crowdsourced content. This presentation will help digital humanists and documentarians spot potential legal and ethical issues, suggest workflows to frontrun potential legal and ethical landmines, and recommend legal forms and specific legal language to shift liability away from a hosting institution and enable scholars to replicate similar crowdsourced community archives in a sustainable, legally-compliant way. Time will also be taken to discuss the implications of AI-generated content submitted through crowdsourced archives that might alter the truth of the historical narrative of a natural disaster. This presentation will also conclude by briefly touching on preservation challenges with crowdsourced archives, including the use of virtual reading rooms and whether that use is compliant with Section 108 of the Copyright Act.

This paper argues that multimodal digital humanities (DH) practices—interactive mapping, network visualization, and digital storytelling—offer powerful ways to revitalize Graham Greene’s The Power and the Glory (1940) and The Lawless Roads (1939) by placing them within the charged political world of the Cristero Era in Post‑Revolutionary Mexico. The 1917 Constitution and the anticlerical intensity of the Calles Era (1924–1928) created an atmosphere in which religious life was monitored, constrained, and often violently suppressed. Greene’s depictions of shuttered churches, persecuted clergy, and displaced rural communities closely reflect these decades of state hostility and the social upheavals that pushed many Mexicans northward. While rooted in a specific historical moment, Greene’s Mexico resonates sharply with contemporary hemispheric patterns. Across the Americas today, religious and social actors continue to face state or regime pressure: Cuba restricts and surveils clergy; Nicaragua and Venezuela have expelled or imprisoned religious leaders; Colombia’s priests contend with violence at the hands of armed groups. These dynamics echo Greene’s attention to communities squeezed between ideology and survival. Of particular interest, this paper draws a structural comparison between Cristero‑era repression and recent federal actions in the United States that have targeted LGBTQ+—and especially transgender—people. Executive orders, federal agency rule changes, and court decisions under the Trump administration have narrowed legal recognition of gender identity, stripped away nondiscrimination protections, and limited access to gender‑affirming care and accurate documentation. Placed alongside Greene’s Mexico, these developments reveal parallel mechanisms where governments police bodies, regulate identity, and define the boundaries of public belonging. Digital humanities methods draw out interesting comparisons and powerful ways to use critical thinking to see through the socio-cultural “reforms”. Map‑based reconstructions will illuminate the geographies of fear and flight Greene recorded; network diagrams will focus on relationships of power, surveillance, and resistance; and short-form digital storytelling invites the audience to pair Greene’s scenes with present-day legislative crises. Such multimodal work foregrounds the voices Greene captures only at the edges—Indigenous communities, rural parishioners, and the itinerant poor—and reframes their experiences within longer histories of cultural erasure and authoritarian control. Ultimately, this project positions Greene’s works as overlooked texts for understanding how oppression is lived, remembered, and contested. Through digital reconstruction and multimodal analysis, the study offers a scholarly and teaching model that is historically grounded and urgently contemporary, demonstrating how literature and digital methods together help us interpret structures of power across past and present Americas.

Interwoven Existence is a kinetic installation that translates environmental wind data into continuous physical motion through a two-axis gimbal driven by an Arduino microcontroller. The work draws from a single day of half-hourly wind measurements from the Lake Taihu Eddy Flux Network (Harvard Dataverse, DOI: 10.7910/DVN/HEWCWM). Rather than simulate meteorological accuracy, the installation renders wind as presence—persistent, subtle, and intermittently still. Forty-eight pre-processed data points function as a fixed score, continuously interpolated into motion. What emerges is not visualization, but embodied translation. The project sits within an ongoing research trajectory in Ecological AI—an approach to computational design that asks how intelligent systems might be shaped by environmental relations rather than solely by human-centered optimization. Within Digital Humanities, Johanna Drucker’s articulation of data as “capta” reminds us that data are constructed and situated. Interwoven Existence extends this insight by approaching wind data not as extractable information to be clarified or controlled, but as atmospheric trace—an archive of lived ecological conditions translated into gesture. Computation here is deliberately material. Echoing N. Katherine Hayles’ emphasis on the physicality of digital systems, the Arduino and dual-motor gimbal make computational processes visible and spatial. The system operates without sensors or audience feedback. It does not adapt or respond. Long pauses interrupt emergent pattern recognition, gently unsettling expectations that technological systems should continually perform or engage. The machine unfolds according to the temporal rhythm of its source data. This design choice opens a speculative question: what if AI systems were not always oriented toward performance, personalization, or efficiency? As Yuk Hui suggests, technologies carry embedded cosmologies. Many contemporary AI systems reflect an instrumental logic—intelligence framed as service. In discussions surrounding General Artificial Intelligence (GAI), intelligence is often imagined as increasingly adaptive, anticipatory, and optimized for human needs. Interwoven Existence proposes another possibility. Ecological AI imagines systems that coexist rather than serve; systems that express persistence, relation, and situatedness. The installation’s autonomy is intentionally modest—a closed dataset, a self-contained mechanism, an unbroken translation of environmental conditions into motion. Yet in this modesty lies a speculative prototype: a suggestion that AI-enhanced cultural systems might cultivate presence rather than productivity. For Digital Humanities, this gesture invites reflection on how emerging intelligent systems participate in cultural ecologies. If DH has long examined how media shape knowledge and interpretation, Ecological AI extends this inquiry toward how design frameworks shape relational worlds. What kinds of futures become possible when intelligence is understood not only as capability, but as mode of coexistence? Previously exhibited at Hanshan Museum (Suzzhou, China) and Ars Electronica (2024), Interwoven Existence offers a contemplative space in which environmental data, machine motion, and human perception meet—inviting ACH audiences to consider how AI might move, quite literally, with the world rather than simply for us.

As AI systems increasingly mediate how human lives are recorded and narrated, the digital humanities field is confronted with questions of interpretation and representational authority. Personal data is predominantly structured to serve institutional, commercial, and administrative interests rather than individual understanding (boyd and Crawford 2012). In marketing, health, and platform economies, narratives generated by individuals are routinely analyzed to predict behavior, segment populations, and optimize decision-making, often without providing meaningful interpretive access to the data producers themselves. Datafication and interface design shape visibility, agency, and power, emphasizing that data representations are never neutral but encode institutional priorities (Drucker 2011; Manovich 2010). Medical humanities is a field within digital humanities that examines how illness, care, and embodiment are experienced and narrated beyond clinical or institutional frameworks (Kleinman 1988; Charon 2006, 2001). In this sense, illnesses are shaped by narrative, temporality, and meaning, and these dimensions are central to how individuals understand their lives over time. Building on this tradition, the project draws on a corpus of publicly available social media data in which individuals narrate their experiences of illness and with medical systems over time through a series of visualizations. This corpus is processed with natural language processing pipelines and large language models (LLMs) to extract thematic labels, temporal anchors, and event sequences from narrative text. References boyd, danah, and Kate Crawford. 2012. “Critical Questions for Big Data.” Information, Communication & Society 15 (5): 662–679. Charon, Rita. 2006. Narrative Medicine: Honoring the Stories of Illness. Oxford: Oxford University Press. Charon, Rita. 2001. “Narrative Medicine: A Model for Empathy, Reflection, Profession, and Trust.” JAMA 286 (15): 1897–1902. Drucker, Johanna. 2011. “Humanities Approaches to Graphical Display.” Digital Humanities Quarterly 5 (1). http://www.digitalhumanities.org/dhq/vol/5/1/000091/000091.html. Drucker, Johanna. 2014. Graphesis: Visual Forms of Knowledge Production. Cambridge, MA: Harvard University Press. Kleinman, Arthur. 1988. The Illness Narratives: Suffering, Healing, and the Human Condition. New York: Basic Books. Manovich, Lev. 2010. “What Is Visualization?” Visual Studies 25 (2): 36–49. Portelli, Alessandro. 1991. The Death of Luigi Trastulli and Other Stories: Form and Meaning in Oral History. Albany: SUNY Press.

This paper examines how contemporary historical video games function as legitimate historical discourse through "Periphery Passive Engagement." Drawing upon frameworks established by Robert Rosenstone and Adam Chapman, I argue that players may acquire historical knowledge through environmental interaction even when bypassing narrative elements such as cutscenes or dialogue. By analyzing the visual and spatial design of games like Assassin's Creed Odyssey, I demonstrate that developers, working alongside professional historians and archaeologists, create historically informed virtual worlds that facilitate passive learning. This research positions intentional historical video games as valid contributions to historical understanding, emphasizing scholarly engagement with this increasingly influential medium.

In this artist’s talk, I’ll discuss a series of poems from Feels Like Power, a manuscript-in-progress that combines the medium of immersive VR with the methodology of documentary poetry. The project explores the ongoing impact of the U.S. military presence in the Pacific region that began during the Pacific War—an emergency that has been “over” for eighty years, but whose aftereffects are still emerging. The poems focus on liminal and contested spaces including perimeter fence zones, war tourism sites, and groundwater drainage systems. They also examine the ways domestic life overlaps with the war machine, allowing a view inside the domestic spaces that house and sustain the thousands of military family members leading suburban American lives on overseas bases in Japan, Korea, Guam, and the Philippines. When you read the poems in 3D, the poems, and their environments, extend panoramically around, above, and below you. By moving your head and body, you can navigate the environment and the text within it. The website enables VR within each poem for those who have VR headsets, but also features 2D versions which are accessible simply with a phone, tablet, or laptop. The project invites you to experience poems spatially in each site, and to travel between sites to discover resonances and contrasts. In their form, the poems recall the popular 19th century panoramic installation paintings of war scenes and distant landscapes a viewer would likely never experience in person. 360˚ panoramas, then and now, promise the confidence of total vision, the ability to see and therefore discover everything available to be known in a scene. But of course they are also merely spheres with finite edges, trapped in the instant the camera shutter clicks. In writing these poems, I hope to leverage immersive VR’s affective power in order to learn about the limits of vision from inside an ideological bubble, the capacity of poetry to see with depth, and the long aftereffects of war as they emerge in our everyday environments and language.

In late 2025, a lab-wide email informed us (the women) that while we were still permitted to speak at the university’s “Women in STEM” event, we would now have to use vacation time to do so. The event violates the federal government’s new anti-DEI rules and our salaries are funded by the Department of Energy. “What’s it like to be a woman in STEM?”, came the joke on our back channels. Amid the scoffs, I caught another sentiment - childish glee. They had forced us underground and coming back from our “doctors’ appointments”, we said nothing. My proposal is an art installation version of those talks in the form of a girls game known as “Sekretiti” (Russian) and “Widoczek” or “Niebko” (Polish). The game appeared somewhat mysteriously across the Soviet Union. Girls would collect and collage small colorful everyday items (candy wrappers, flower petals, beads) then place them beneath a piece of recycled glass in a small hole dug for that purpose. Later, they would share their treasure-window among friends, in secret. While the U.S. government goes through a bombastic renaissance of federally-sanctioned archival silencing, digital information professionals and their extended communities are working to track and “rescue” data. It’s undoubtedly important work. Yet, there’s a nagging question: “Who should we share our information with?” In the creative game of Sekretiki/Niebko, is something that can interrupt our digital, immediacy-driven instincts to preserve and share loudly. I’m interested in uses of initial secrecy to strengthen digital privacy. Secreteki can be uncovered, but it requires some effort; not just a click, but some labor, or learning, or trust.

As Hannah Ringler lays out in her chapter “Computation and Hermeneutics” (Computational Humanities, ed. Johnson, Mimno & Tilton, 2024), we can think of developing interpretive practices for computational methods as “asking questions with and about tools.” In the context of the large language models as an emergent tool, fostering critical understanding of how they work and their relationship to human creativity is a teaching imperative. This lightning talk/poster will share the results of a pedagogical experiment in orienting an undergraduate research team of 4 students to the critical practice of computational literary studies by prompting them to ask questions with and about large language model text generation by asking questions with and about Gertrude Stein’s poetics. This experiment takes the form of a three-part opening module for a 10-week term of undergraduate research. These three parts are: primary source reading, computational experimentation, and reflection. Primary source reading (selections from Stein's poetry collection Tender Buttons and her essay "Portraits and Repetition," in which she talks about writing these poems) immerses students in an experience of linguistic defamiliarization. After initial readings, students are asked to select a specific poem to focus on and attempt to specify what linguistic patterns seem emphasized in the poems and then make an argument about whether these patterns can be linked to specific propositions in Stein's stated poetics. The computational experimentation phase begins by asking students to read a chapter on text generation using a transformer-based large language model and then work through the accompanying Jupyter notebook to familiarize themselves with parameters including temperature and n-gram tokens. Students are then asked to conduct experiments to attempt to find settings that will get them as close as possible to generating their own Tender Buttons-style short poem. Finally, students compare a selected object poem to the "closest" generated text and make a preliminary argument about whether Steinian aesthetics could ever be machine-generated (which would entail arguments about how to define both the aesthetics and the measurement of the success of machine-generation). At this time, I am able to share the module outline. By the time of the conference, I will be able to report on its implementation, and I anticipate being able to share reflections on questions including the effectiveness of Steinian poetics as the central textual point of inquiry; specific technical pain points for students from a literary background and reading pain points for students from a computing background; growth in critical understanding of LLM mechanics; and how this module shaped students’ intentional and critical practices of interpretation of computational outputs.

Since Zygmunt Bauman’s theorization of liquid modernity, identity has been understood as both constructed and unstable—a site of emergence that simultaneously constitutes a state of emergency (Bauman 1996). In literary criticism, the question of authorial identity—its definition and circumscription within the text—has long anticipated broader debates on the instability of subjectivity. The role of style as a means of individuating the author can be traced back at least to the seventeenth century (Compagnon 1998). It is therefore unsurprising that stylometry—the statistical analysis of literary style—has found its most consistent application in authorship attribution (Holmes & Kardos 2012). This paper illustrates a multivariate method for authorship attribution that integrates computational analysis with literary interpretation. Beyond identifying probable authors, the method aims to illuminate writing practices and the layered construction of authorial identity within a collaborative text.
The case study is Q (1999), a historical novel published under the collective pseudonym Luther Blissett by four Italian authors: Roberto Bui, Giovanni Cattabriga, Luca Di Meo, and Federico Guglielmi. Reconstructing the violent aftermath of Luther’s Ninety-Five Theses—with ecclesiastical spies and brutal war scenes between soldiers and Anabaptist subversives—the novel becomes a distorted mirror of our own time, marked by the return of conservative ideologies, and a long allegory about history’s alternative possibilities. Only three of the authors produced solo works, which form the reference corpus for this study. To date, no systematic attempt has been made to attribute individual sections of the novel. The text was divided by a native Italian literary scholar into units smaller than chapters, marked by “***”, resulting in 232 sections. The aim was to attribute each section to one of the three authors with available solo publications. The multivariate method combined function-word and content-word analyses. Function words, often considered reliable unconscious markers of style, were examined through principal component analysis (Craig 2009–10), which identified four highly discriminative items (una, che, non, and di). Burrows’s Delta (Burrows 2002) was then applied to measure stylistic distance between sections of Q and the authors’ solo works, producing probabilistic attributions. Content-word analysis provided complementary results. Average sentence length (Hjort 2007) revealed stylistic polarization: one author consistently employed longer sentences (over 20 words), while the others favored shorter constructions (7–20 words). The recurrent use of specific lexical fields, particularly swear words, further distinguished one author within the latter group. Notably, function-word and content-word analyses did not consistently converge. This discrepancy prompted a reconsideration of the novel’s compositional process. In a 2019 interview, Giovanni Cattabriga described a collaborative method in which chapters were drafted collectively and subsequently revised—or entirely rewritten—by another member. The presence of dual stylistic signatures within sections thus reflects the group’s layered writing practice rather than methodological inconsistency. 
These findings underscore both the stratified nature of collaborative authorship and the slipperiness of authorial identity, further complicated by collaborative authorship practices. Like identity in liquid modernity, the authorial self emerges through overlapping intentional and unconscious traces, revealing writing as a dynamic and multi-layered process.

From 2024-2025, we engaged in an undergraduate research project focused on digital reparative archives, maker culture and zines (Reparative Archives Zotero Group, 2024), and the cultural implications of Artificial Intelligence. A central question continued to resurface: what will happen to humanity’s sense of community as AI locks individuals into solitary digital bubbles? AI’s affordances allow it to be a surrogate for human connection, drastically altering humanity’s trust in technology, cultural thinking, and social media. As AI expands into more use cases across domestic and professional life, we must evaluate AI’s potential impact before it dominates our engagement within communities. This creative presentation will showcase a zine titled “Dimensions Colliding: We Asked for a Third Space and All We Got Was This AI-Generated T-Shirt”, which was developed as a culmination of several years working with the Lincoln Center for Applied Ethics at ASU. The zine will contain sections such as: (1) summarized contextual research for public-facing audiences and (2) direct quotations from interviews with the public in a collage style framework. In addition to showcasing the physical zine, we will present the digital version of the zine, to be published online using minimal computing principles. This project defines “digital social climate” as a collision of two distinct dimensions. First, social climate refers to shared perceptions of beliefs about a social environment held by a group of people (Bennett, 2010). Second, digital climate refers to a space where people’s shared perceptions, attitudes, and behaviors towards digital technologies are shaped by digital infrastructures (Avtalion et al, 2024). When these two dimensions collide, they create what we call the “digital social climate”, a newly emerged third space where perceptions or beliefs can reconfigure how people interact, collaborate, engage, and innovate. AI’s intrusion into this third space poses significant challenges to our ability to connect meaningfully with each other in digital space. AI erodes community, especially as we approach the reality of the dead internet, where most of the content we see is AI-generated (Tiffany, 2021). By producing a zine in both physical and digital forms, this project will engage with the metaphor of colliding dimensions to resist the use of AI as creator/artist of media (examples of this include Tilly Norwood, the AI-generated “actress”, or Xania Monet, the AI-generated "musician" with 1.3 monthly Spotify listeners). By asking conference goers to engage in play within this third dimension, especially using creative expression as a way of communicating with and activating community, we will use the methods of #DHMakes to further interrogate one of the original divides in digital humanities: make/hack vs. critique/yack (Visconti, et al., 2024). Presentation language will be primarily English, but also bilingual. Avtalion, Z., et al (2024). Digital Infrastructure as a New Organizational Digital Climate Dimension. Applied Sciences. Bennett, J.B. (2010). Social Climate Research. Tiffany, K. (2021). Maybe You Missed It, but the Internet “Died” Five Years Ago. The Atlantic. Visconti, A. , et al. (2024). "#DHmakes: Baking Craft into DH Discourse." Korean Journal of DH. Reparative Archives Zotero Group. (2024). https://www.zotero.org/groups/5583269/reparative_archives

As the hype around artificial intelligence continues to shape everyday life, this creative presentation shares a practical and student-centered way of teaching AI literacy, especially large language models (LLMs), through a digital humanities course. This DH course treats AI both as emergence (new literacies) and emergency (environmental and infrastructural stakes). Designed for undergraduates across humanities and social science backgrounds, it combines concept learning, a data center field visit, language model testing, and collaborative digital project building with AI assistance. This pedagogy experience is distinct yet replicable. This course is framed in response to a recent systematic literature review of DH pedagogy (2012-2022), which points out that critical humanistic analysis of digital resources and infrastructure remains rather rare (Georgopoulou et al.). In response, this course begins with care through conceptual clarity, so students can engage confidently in AI-related conversations. They distinguish key terms such as AI, machine learning, deep learning, neural networks, generative AI, and language models. They then create accessible glossaries and explainers for their peers. Rather than immediately turning to commercial LLMs, students work with small and medium-sized language models in LM Studio on the university’s high-performance computing (HPC) resources. This hands-on testing reveals how model size shapes performance and how prompts could trigger phenomena like hallucination. Such experience sets the stage for deeper critical engagement. The course then shifts to emergency by discussing the environmental implications of AI. Through a field trip to the university’s own data center, students feel the heat firsthand, observe water usage, and witness the physical infrastructure behind AI. Guided discussions on energy and water consumption, e-waste, and mining impact help students recognize that AI does not exist simply in the “cloud,” as mainstream narratives suggest, but as a system embedded in ecological and community contexts. In the second half of the course, students collaboratively design a sentiment analysis project using media coverage of data centers in Michigan. They learn traditional sentiment analysis workflows (collecting articles, cleaning text, and using sentiment analysis tools) alongside AI-assisted approaches. The class compares four pipelines: 1. generating code with a commercial large language model and running it locally; 2. generating code with a small/medium model and running it locally; 3. feeding cleaned text directly to a commercial large language model for code‑free sentiment analysis; and 4. sending text to a small language model to classify sentiment through language understanding alone. Students analyze differences in accuracy and nuance while reflecting on the methodological trade-offs of each approach. Findings are then translated into public‑facing resources: a collaboratively built GitHub Pages site and an academic poster. This course is designed around care (conceptual literacy and peer education), response (hands-on model testing and data center site visit), and transformation (public scholarship); it hopes to strengthen DH pedagogy in a time heavily impacted by rapid AI expansion. As the course is currently running in Spring 2026, the presentation will provide links to both the poster and the website later.

We are two digital humanities workers who engage in feminist praxis, relying on friendship as our organizational framework to refuse the pressures of extractive neoliberal productivity. By resisting what we call the higher ed managerial industrial complex (developed from Ginsberg’s “All-Administrative University,” 2011), we actively subvert the tools of modern academic management, replacing corporate “care” with care-full collaboration. Drawing on the work of Saidiya Hartman and Tina Campt (Refusal Collective), Bonnie Honig (2021), adrienne maree brown (2019), Whitty and Wheeldon (2025), as well as cooperation theory (Axelrod, 1984) and projects like the Feminist Data Manifest-No (2019), we approach our work from a place of refusal—a political and strategic stance to resist the managerial tools and pressures being thrust upon us. We strive for a human-centered approach to our collaborations—an approach that insists that meaningful, values-infused work is not possible without friendship. We are in the process of developing a manifesto that uses our friendship as an experimental lens to explore ways that we, as digital humanists, can resist the managerial structures and contexts within which we undertake our work. If we take up friendship as a DH methodology, how can we then unravel the interconnected impulses of DH as an institutionalized entity within the neoliberal university, and in contrast, envision DH as potentially radical and liberatory? This creative presentation documents the conditions that make anti-managerial cooperative relationships possible. Our project grows out of and through a series of conversations between two friends who found each other through the Association for Computers and the Humanities, where we now serve as Co-Vice President/Co-President Elect. We propose that dialogues—multimodel explorations of written, spoken, and woven laughter, tears, and words—are a generative pathway that helps us trace our individual feminist lineages as they ran parallel, eventually intersected, and finally intertwined. These autoethnographic explorations help us to imagine new possibilities for liberatory, anti-management DH labor. An ongoing project, our creative presentation will provide an extra-institutional space to open up questions for consideration, suggest exercises for intention-setting, and explore techniques to divert the incessant/insatiable demands of the institution. We do not offer a set of best practices; rather, we propose an ongoing and always evolving practice/praxis of refusal and resistance within and beyond the confines of our respective neoliberal institutions. We see this creative offering as directly addressing ACH2026’s call for work that speaks to the collective power of community to forge solidarity in times of great crisis. Just as we struggle against the neoliberal university’s extractive managerial methods, our institutions are struggling against the pressures of the current federal administration, threatening our work and livelihoods. In practicing friendship-as-method for digital humanities, we hope to locate lifelines for the ACH community to chart their own survival mechanisms.

From textiles as a medium for data visualization, to 3D printing whistles as a rapid response action, to letterpress offering an alternative vision for what it means to create text in an era of AI, to celebrating craft-making as a hobby that offers a path to connect to materiality in a digital world, #DHmakes has brought together a broad, distributed community of makers engaging with these themes (Visconti et al 2024). For the past several years, members of the #DHmakes community have organized a virtual-world hangout space for makers to gather and share their works in progress at the ACH conference (Dombrowski 2026). For ACH 2026, we would like to theme a gathering around the idea of a virtual neighborhood picnic: coming together “in person” with strangers, and building relationships through sharing the things that we’re excited about. The spirit of this event is intended to echo the community organizing work happening within local neighborhoods in response to the threat posed by government agents, with a similar goal of fostering networks of care and future collective action that can support participants moving forward. We will provide a sign-up sheet for conference attendees a week in advance, and it will remain open until the morning of our presentation slot. “Drop-ins” will be welcome if anyone who has registered to present is unable to come at the last minute. Participants will have 2 minutes to share something they’re making, or have recently made. We’ll plan on up to 15 2-minute presentations, followed by 30 minutes of “co-making” time, where participants are invited to work on (some aspect of) their craft at the same time, leaving space for spontaneous conversation. As a final component of this activity, and to accommodate people who are unable to attend during the synchronous time slot, we will invite all conference participants to submit a 6" square image of something they’ve made, along with a few sentences describing it. Organizers will compile these into a single “picnic blanket” image and publish it online along with the statements, and make a high-resolution version available for participants to print (e.g. as a poster, t-shirt, or basis for further craft work) as they would like. References Dombrowski, Quinn. “Craftivism in a Crisis: Making the Humanities Matter When It’s All Falling Apart”. CESTA seminar, January 29, 2026. Writeup at https://quinndombrowski.com/blog/2026/02/10/craftivism-crisis/. Visconti, Amanda et al. “#DHmakes: Baking Craft into DH Discourse”. Korean Journal of Digital Humanities, vol. 1, issue 1. 2024. https://doi.org/10.23287/KJDH.2024.1.1.5

This study situates Indian electronic literature within the broader ecosystem of India’s rapidly expanding creative industries through a comparative, interdisciplinary, and context-sensitive framework. Drawing on recent policy and industry reports, most notably Arts and Technologies in India: Reimagining the Future (Barua et al., 2024) and PwC’s From Sunrise to Sunshine (2024), it examines why digital art and video games in India have achieved far greater visibility, institutional support, and cultural traction than electronic literature, despite all three being grounded in computational and networked technologies. While electronic literature emerged early as a global practice and has developed robust communities in Europe, North America, and East Asia, its growth in India has been comparatively slow and fragmented. By contrast, digital art and video games have thrived by aligning with broader economic narratives, audience engagement strategies, and policy frameworks. Barua et al. (2024) highlight how Indian digital art has expanded beyond traditional gatekeepers through interdisciplinary practices, hybrid media forms, and international collaboration, even while facing infrastructural and access challenges. Similarly, PwC (2024) documents the rapid expansion of India’s gaming industry, driven by localization, indie studios, governmental initiatives, and integration into national development agendas such as “Vikshit Bharat.” Existing scholarship on Indian electronic literature identifies barriers such as the digital divide, limited infrastructure, and lack of awareness. While these factors remain crucial, this paper argues that they are insufficient without a comparative analysis of adjacent creative industries that have successfully navigated similar constraints. Electronic literature’s marginalization, the paper contends, stems less from inherent limitations than from its exclusion from creative-industry policy narratives and its continued confinement within print-centric and Western academic paradigms. The study proposes reimagining electronic literature as part of a continuum of interactive, narrative-driven digital practices. It concludes by advocating the establishment of the Indian Consortium for Interactive Digital Narratives (ICIDN), a cross-sectoral platform designed to foster collaboration, institutional support, and sustainable growth, thereby securing electronic literature’s place within India’s evolving creative economy. References: Barua, K., Rishi Gokharu, Afrah Mutaher, Nupur Singh, Hannah Andrews, Ruchi Das, Andrew Hawcroft, Radhika Rao, Aastha Sodhani, and Divya Sundara Raja. Arts and Technologies in India: Reimagining the Future. British Council, 2024. Ensslin, Astrid, Samya Brata Roy. “Electronic LiteratureS as Postcomparative Media.” CompLit: Journal of European Literature, Arts and Society, 2023, pp.145-171, Classiques Garnier. Mukherjee, Souvik. “No Country for E-Lit?--India and Electronic Literature.” Hyperrhiz: New Media Cultures, vol.16, 2017. PricewaterhouseCoopers. From Sunrise to Sunshine: The Contribution of Online Gaming to the Viksit Bharat Journey and India’s Cultural Power. PwC India, 2024. Roy, Samya Brata. “e-Sahitya, or the Contested Emergence of the Electronic within the Literary in India.” The New Review of Hypermedia & Multimedia, vol. 27, no. 3, 2025. Shanmugapriya T., Nirmala Menon. “First and Second Waves of Indian Electronic Literature.” Journal of Comparative Literature and Aesthetics, vol.42, 2020, https://jcla.in/wp-content/uploads/2020/10/JCLA-42.4_E-Lit_Nirmala-Menon.pdf.

Recent advances in AI enable large-scale digitization and post-processing of oral history collections, including automatic speech-to-text transcription, time-stamping, annotation with Large Language Models, and multimodal video analysis, e.g., facial features or inferred stress indicators (Rønningstad et al., 2024; Yakaew et al., 2021). While these methods expand what can be computed from audio/video at scale, less attention has been paid to how domain experts—such as historians and cultural memory studies scholars without computational training—can meaningfully work with AI-produced outputs. A central practical challenge remains: digital platforms can filter and facet large collections, yet the resulting sets are still too large for close listening and watching. This presentation demonstrates how data-driven selection (specifically, a regression-informed selection) can support the identification of analytically useful subsets of interviews, helping human experts engage with computationally processed oral history at scale while reducing the risk of invisible selection bias. We present a case study from the “Voices from Auschwitz” project, which investigates survivor testimonies to contribute to the history of Auschwitz from victims’ perspectives. The project combines computational post-processing (e.g., transcript time-stamping, facial analysis) with qualitative analysis by domain specialists (including a perpetrator historian and a cultural memory scholar). The focus of our case study is on interviewer–interviewee dynamics, widely recognized in oral history scholarship as shaping narrative form, disclosed content, and remembered experiences (Lang et al., 2023, Pattinson, 2011). Interviews can vary from interviewer-dominated exchanges—where information must be actively “pulled” through frequent questioning—to interviewee-dominated narratives with few prompts, as well as many intermediate interaction patterns. The dataset of our case study comprises 1,000 English-language Holocaust testimonies from the USC Shoah Foundation, transcribed using a deep learning model. The transcripts distinguish interviewer questions and interviewee answers with timestamps. We model interaction dynamics using two variables: interview length (X-axis) and total number of questions/answers (Y-axis). A scatter plot (Figure 1, not uploaded since Conftool supports only pasting) positions each interview as a point; interviews with relatively few questions (more interviewee-driven) fall near the X-axis, while highly question-driven interviews appear higher on the Y-axis. A regression line estimates the expected number of questions given interview length. To curate a manageable set for close analysis, we use sampling based on residual distance from the regression line (Williams, 2013). We (1) compute residuals for all interviews, (2) estimate the residual distribution, and (3) draw random samples from this distribution, matching each draw to the closest real interviews. We also sample from the distribution’s tails to capture outliers. This yields a subset of 50 interviews representing typical cases (near the regression), contrasting dynamics (well above/below it), and extreme patterns. We argue that this regression-informed selection process offers a principled alternative to ad hoc selection: it broadens coverage of interaction types, reduces inadvertent bias, and helps domain experts productively navigate the “too much to watch” problem in AI-enriched oral history collections.

The field of digital humanities is increasingly being influenced by the current environmental crisis which has contributed to the emergence of an interdisciplinary field merging environmental humanities with digital humanities, the Digital Environmental Humanities (DEH). Scholars of DEH often grapple with a seeming contradiction: while the terms "digital" and "environment" appear almost contradictory, DEH acknowledges the environmental damage inherently caused by the digital humanities field itself. Beyond this focus, the discipline also explores how digital methods can be utilized to address and mitigate environmental loss. Examples include regional digital labs in the US dedicated to sustainability issues such as drought and water rights, and public resources like the open-access Native American Ethnobotany Database (Ryan, Hearn, and Arthur 2023). Drawing on the Digital Environmental Humanities (DEH) framework, this paper addresses the ACH 2026's central theme of emergence/emergencia. It poses a critical question for digital humanities scholars: how can our knowledge creation effectively critique ecological emergencies without inadvertently reinforcing the very conditions we aim to challenge? Our research proposes an eco-computational comparative analysis of Indian ecological drama and American eco-theatre, examining their approaches to the intensifying crisis of climate change. American eco-theatre provides a strong basis for this study, characterized by both canonical works and grassroots efforts that have historically heightened environmental awareness, tackled specific local ecological issues, and translated complex scientific data into relatable human stories (Cless 1996; May 2005; May 2020). Conversely, Indian drama has consistently given prominence to the environment since its origins, frequently utilizing natural settings as a pivotal, recurring motif that often acts as a key catalyst, both instigating and reflecting the emotional experiences of its characters (Forthcoming). Thus, the central methodological inquiry guiding this study is: Considering these distinct geographical and dramatic traditions, how is the complex reality of environmental emergency staged, interpreted, and communicated as an experience that is lived, embodied, and fundamentally interconnected? To answer this, the study applies an 'eco-computational' approach by employing digital humanities tools such as topic modeling, network analysis, and sentiment analysis. This computational approach will reveal patterns and structural differences less obvious through traditional close-reading, thereby deepening the understanding of crisis rhetoric, human-nature relationships, and ecological justice. The comparison aims to shed light on shared global anxieties and divergent cultural responses to the Anthropocene. Keeping environmental sustainability as a core principle, we aim at deploying low-carbon digital methodologies, such as the analysis of text samples of reduced size, the systematic mapping of salient environmental terminology, and an examination of the co-occurrence of narrative agents and their respective environments. The approach transcends a mere investigation of the environmental crisis by positing the crisis itself as the epistemic catalyst guiding the research execution. This study also represents a re-evaluation of Digital Environmental Humanities (DEH) to emphasize computational resource prudence and ethical consumption. Furthermore, this work introduces the concept of Indian Digital Environmental Humanities (IDEH), which is a regionally rooted yet transferable framework that offers methodological insights relevant to digital environmental scholarship, including DEH scholarship emerging across the Americas.

Language has always existed within material infrastructures. Each communicative technology - oral performance, manuscript, print, and now the digital - reshapes not only how we transmit words but how we know. This paper traces the history of South Asian communcation to ask how digital and AI-based systems might recover the pluralism that print once constrained. Pre‑print South Asia operated through fluid multilingual and multiscript worlds, where texts moved easily across Sanskrit, Persian and vernacular registers. Scripts were interchangeable and variation was integral to meaning. The arrival of colonial print introduced fixity: languages were tied to specific scripts, orthographies standardized, and linguistic boundaries hardened into instruments of governance. Print’s epistemic order produced a new ideal of purity, erasing the manuscript era’s tolerance for variation and collective authorship. Today, digital technologies place us again at a threshold of emergence and emergency. Early computing repeated print’s discipline using ASCII, Unicode, and the typographic standardization of scripts, but deep learning has begun to unsettle these limits. Neural networks trained on irregular, noisy and hybrid data can now read handwritten manuscripts, complex Indic scripts and degraded archives, reopening access to multilingual histories long excluded from the digital record. Yet these technical possibilities raise ethical and social questions: What languages are prioritized for digitization? Who participates in training and annotation? Can AI systems embody the pluralism they help recover? Grounded in case studies of digitization projects across Indian languages, this paper argues that the digital turn can either reproduce colonial hierarchies or enable new solidarities. To design for plurality, we must treat imperfection, incompleteness, and multilinguality not as problems but as sources of knowledge. The digital humanities thus emerge as a space of care where technological design and cultural memory converge to rebuild a more fluid, inclusive linguistic infrastructure for the future.

Digitisation often happens in a hurry, where materials are fragile, teams are small, funding windows are short, and audiences expect online access now. In that pressure, GLAM projects increasingly turn to OCR and, more recently, to generative AI to speed up description, transcription, and search. But in multilingual and multiscript collections, these tools can widen the gap between ‘being visible online’ and ‘being findable and usable’ because items may be 'digitized' while still difficult to search, interpret, or reuse, especially for communities working in non-dominant languages and scripts. Framed through ACH 2026’s theme of emergence/ia, this paper argues that our most urgent need is not only better automation, but clearer, as well as more shared ways of explaining what automation ‘can, and cannot, do’. The paper offers a policy- and guidelines-based critique of current digitisation and AI-use practices in cultural heritage, where the focus is on a recurring problem- institutional documentation often says too little about multilingual coverage, error behavior, and who is accountable when automated outputs fail. Across common forms of guidance, we repeatedly see familiar defaults which include- English-first metadata expectations, inconsistent language or script representation, and broad claims such as ‘searchable’ or ‘AI-enhanced accesses’ without any explanation of scope or limits. These are not just technical details, but they shape who can find materials, what kinds of research become easier, and whose records circulate as credible sources, which raise questions about equity, labor, and transnational knowledge exchange. As a practical contribution, this paper proposes two lightweight artifacts for small teams operating under constraint. First, a ‘Do-Not-Overclaim Checklist’ for OCR or LLM-assisted projects that codifies what projects should report about language/script coverage, evaluation method, uncertainty, and known failure modes. Second, a ‘Multilingual Discoverability Statement template’ that can be appended to project pages or catalog entries, allowing users to rapidly grasp what is known, what is partial, and what is not to be assumed. In centering multilingual standards and accountable reporting, this paper contributes to a solidarity-oriented approach to DH infrastructure development that makes access claims transparent and comparable, facilitates transborder research communities, and leaves room for later empirical pilots without making them a requirement at the proposal stage.

The study of ephemeral cultural phenomena, such as wartime exhibitions, is often hampered by fragmented and degraded archival records, pushing the limits of traditional historical inquiry. This paper addresses this challenge by detailing the theory and practice of an interdisciplinary methodology, termed “digital historical forensics,” that fuses computer vision, archival historiography and exhibition studies to reconstruct such lost visual experiences. We present a case study focused on photographic exhibitions in WWII China (1937–1945) to demonstrate the mechanisms and processes required to make interdisciplinary research effective in a real-world scenario where source data is sparse. Our methodology integrates three distinct phases: First, we employ deep learning techniques, including latent diffusion models for super-resolution, to enhance low-quality historical images. Second, we introduce a novel image search algorithm, Hierarchical Region Pursuit (HRP), that leverages a transformer-based matcher to forensically identify displayed photographs by cross-referencing them against a large archival database. The third, and most critical, phase is an iterative process of historical contextualization, where computational outputs are critically vetted and interpreted by human experts using primary sources and secondary literature. This workflow exemplifies an integration of disparate disciplinary approaches, where computational findings can challenge established historical chronologies, prompting new avenues of humanistic inquiry. Applying this framework, we successfully reconstructed significant portions of two exhibitions from 1940s China, recovering the specific images displayed. This enabled a granular analysis of spatial layouts, curatorial strategies, and inferred viewing cues of historically significant exhibitions that would otherwise remain speculative. This case study serves as a critical reflection on the practice of interdisciplinarity, highlighting its value in generating knowledge unattainable by a single discipline. Ultimately, this research offers a robust and scalable model for transforming scant data into rich, interpretive insight for future historical research.

The Recetas de las Américas project is a digital humanities project based on a dataset of historical newspaper recipes published in Diario las Américas (Miami, Fla.) from 1953 to 1963. In this presentation, the creators of this project will present the newspaper recipes, discuss our workflows and challenges in data transformation, and share preliminary findings from early analyses of the dataset. About the Recipes Still in print today, Diario Las Américas debuted in 1953 as the first Spanish-language newspaper published in South Florida. The newspaper’s early tagline “por la libertad, la cultura y la solidaridad hemisferica” (“for liberty, culture and hemispheric solidarity”) reflected its Pan-American perspective and dedication to shared cultural life and exchange in Miami and abroad. Although primarily published in Spanish, the last page of each issue was dedicated to English-language sections and bilingual language lessons. Published in Miami, FL during a period of rapid, significant demographic change driven by migration from Latin America and the Caribbean, these recipes are both distinctly Latin American and deeply local. They appear in the “Del Hogar” (“Of the Home”) section of the Diario, a page oriented toward women and homemaking where readers could find the latest fashion trends, advice columns, and domestic guidance including recipes. Collected and selected by Victoria Alejandra, section editor of "Del Hogar", the recipes bear witness to the daily lives, domestic labor, and cultural continuity among immigrant communities navigating resettlement. Written almost entirely in Spanish, with occasional English brand names, and substitutions offered for ingredients not as easily found in North America, these recipes not only function as an informal archive of memory but also highlight inter-linguistic and inter-cultural realities of the time. About the Project The Recetas de las Américas project [https://recetas.domains.uflib.ufl.edu/] seeks to reconnect present-day Floridians with the past through food and culture. The project currently makes 54 recipes available for browsing, searching, and printing available online. This presentation will focus on recent work to expand the dataset to the full 1,076 recipes published from 1953 to 1963 in issues of Diario Las Américas held at the University of Florida and digitized through the National Digital Newspaper Program [https://www.loc.gov/item/sn82001257]. This presentation will not focus on the addition of these recipes to the website, but rather the pre-processing necessary for web publication. About the Presentation In this talk, we will present the dataset alongside the challenges we faced in making it and the workflows we developed to meet those challenges while maintaining the integrity of the recipes. In this presentation, we will focus on the natural language processing (NLP) required, framing our work as a comparative analysis of NLP using traditional human-only coding, AI-assisted coding, direct AI translation, and automatic (non-AI) translation in Microsoft Excel. The presentation will also discuss preliminary analysis of the full dataset, including patterns in ingredients, dish types, publication frequency, formats, and sources to consider what these recipes reveal about Latin American foodways, cultural identity, and everyday life in Miami during the 50s and 60s.

The Latinx Catskills (discover.latinxcatskills.com) is an early-stage public, digital humanities project that digitizes, documents, and preserves the longstanding though effaced presence of Latinxs in New York’s Catskills region from the late nineteenth century through the 1970s. While this history is nearly forgotten today, the Catskills, located 100 miles northwest of New York City, was once a popular summer destination for countless Latinxs. For nearly one hundred years, Latinx travelers took day trips, stayed at resorts owned or managed by fellow Spanish speakers, indulged in the lively music scene, and enjoyed the region’s opportunities for leisure and outdoor recreation—practices which shaped Latinx social life in the Northeast and helped establish what we have conceived of too often as exclusively urban identities. Further, as this project shows, the region—once commonly known as the “Spanish Alps” and later dubbed the “Puerto Rican Alps”— provides a significant geographical space for understanding the cultural productions of Latin American, US Latinx, and Spanish figures, from José Martí and Federico García Lorca to José Juan Tablada and Tito Puente. Our project, which is being implemented in CurateScape/Omeka, raises critical questions about the place of recreation and leisure practices in the archives and the role of the rural retreat in US Latinx experience. US Latinx lives have often been understood primarily in relation to work and productivity. The Latinx Catskills reconsiders New York Latinidades beyond and outside the spheres of urban toil. Scholars of “Black joy” and thinkers like Audre Lorde and bell hooks have shown that care, leisure, and emotional plenitude can be forms of resistance against racialized capitalism. Similarly, the Latinx Catskills archive positions recreation and enjoyment as critical dimensions of diasporic life—especially for communities historically defined by exploitation. By investigating the Catskills region as an affective and social infrastructure for the Latinx good life, this paper, and the project on which it is based, contributes to the transformative momentum that the Latinx Digital Humanities has had in showcasing the complexities of Latinx identities while challenging narrow definitions of Americanness (cf., inter alia, Baeza Ventura, López, Cotera). The digital recovery and dissemination of archival materials can help reclaim marginalized Latinx experiences, highlighting alternative narratives of belonging and resistance in the United States. Aligned with the conference's commitment to interrogating, and responding to, narratives and moments of emergency, emergence, and emergencia, this paper emphasizes DH praxis as an urgent means of documenting the richness of Latinx experiences, remapping both cultural and geographical landscapes.

Esta ponencia examina Archivo Venezuela y Archivo Studio como respuestas situadas a la emergencia y lo emergente en las humanidades digitales de las Américas. Ambos proyectos surgen en el contexto de la crisis de desplazamiento venezolano (la mayor migración contemporánea del hemisferio occidental) y abordan una pregunta central para ACH 2026: ¿cómo pueden las humanidades digitales crear infraestructuras de conocimiento que respondan a situaciones de urgencia y, al mismo tiempo, posibiliten nuevas formas de colaboración, memoria y solidaridad transnacional? Archivo Venezuela es un archivo digital transfronterizo que documenta la producción cultural de la diáspora venezolana: literatura, artes visuales, performance y creación nativa digital. Más que un repositorio neutral, lo entendemos como una infraestructura de cuidado, configurada por condiciones específicas: migración forzada, dependencia de plataformas comerciales, asimetrías lingüísticas y precariedad material. En este contexto, los objetos culturales circulan por ecologías digitales inestables (enlaces rotos, plataformas efímeras, infraestructuras nacionales desiguales) que vuelven especialmente vulnerables las memorias diaspóricas. Un eje central del análisis es Archivo Studio, una interfaz modular asistida por inteligencia artificial diseñada para sostener el archivo bajo estas condiciones. Archivo Studio media entre plataformas comerciales y sistemas de preservación de largo plazo mediante flujos de trabajo automatizados: extracción de metadatos, descripción bilingüe, control de calidad y generación de respaldos. La inteligencia artificial no funciona aquí como herramienta predictiva o extractiva, sino como un mecanismo de triaje infraestructural que redistribuye el trabajo curatorial, apoya la producción de descripciones multilingües y permite mantener registros culturales a escala en contextos de recursos humanos e institucionales limitados. Desde los estudios archivísticos, la teoría crítica de infraestructura y los estudios de la diáspora, analizamos cómo condiciones de emergencia (desplazamiento, inestabilidad tecnológica, ausencia institucional) dan lugar a prácticas emergentes en las humanidades digitales. Estas incluyen estrategias de preservación preventiva, diseños orientados al mantenimiento y decisiones éticas sobre visibilidad, autoría y participación comunitaria. Examinamos el uso de automatización en Python, esquemas de metadatos bilingües basados en Dublin Core y la producción de manifiestos de preservación como prácticas de conocimiento moldeadas por la urgencia y el cuidado. Al centrar un proyecto arraigado en la experiencia diaspórica latinoamericana, esta ponencia responde al llamado de ACH a descentralizar las narrativas estadounidenses sobre innovación en humanidades digitales. Sostenemos que los archivos no son casos periféricos, sino espacios críticos donde la emergencia produce nuevas formas de imaginación infraestructural.

This project proposes the creation of an interactive digital dictionary of fifteen lunfardo terms that are widely used in Argentina today. Lunfardo is the slang of the Río de la Plata region (Buenos Aires and Montevideo), originated in the late nineteenth and early twentieth centuries and shaped primarily by the daily experiences of Italian immigration. Many of its terms became popular through tango lyrics, and, over time, entered common speech, where they continue to evolve in meaning and usage. Designed for college-level Spanish students, general Spanish speakers, and lovers of languages, each dictionary entry will present the lunfardo word, its standard Spanish equivalent, an English translation, sample tango lyrics, links to the song and video, and an explanation of its cultural significance. Some examples are “morfar” (“to eat”), “mufa” (“bad luck”), a word that was very much used during the 2022 FIFA World Cup, and “falluto” (“fake”, “dishonest”). In addition, each entry will provide a link to current songs, news articles, comics, or videos, that demonstrate contemporary use of the term, accompanied by reflections on the historical evolution and cultural implications of the term. This exhibit responds to a cultural and linguistic emergency: the risk that regional languages such as lunfardo may be stereotyped, detached from their historical roots, and simply used for commercial interests as they circulate through global media and digital platforms. Tango is internationally recognized as a symbol of Argentina, but the linguistic richness embedded in its lyrics is often overlooked. In this context, registering and contextualizing lunfardo, emphasizing its local use, becomes a form of cultural preservation and dissemination. The project aims to include current songs, videos or comics, which can function as reflections of authentic uses of the language. Lunfardo emerged from transnational movements. Its origins date back to 1870–1880, during the period of massive European immigration to Argentina, particularly from Italy. In its early decades, approximately half of lunfardo’s vocabulary came from Italian dialects,These terms blended with popular Spanish, African-derived words (many arriving via Brazil), Indigenous languages, and rural expressions. By tracing these histories, the exhibit situates lunfardo within broader processes of migration and cultural exchange. At the same time, contextualizing lunfardo in its current local usage as well as in its transnational reception, shows how living languages are the fruit of continuing contact and relations among humans. By building the project as a digital exhibit using the Omeka platform, the exhibit demonstrates how digital humanities methods can create new ways of registering and communicating linguistic change over time. Rather than functioning as a static dictionary, the project will serve as a curated, interpretive space that highlights connections between language, music, history, and national identity. It shows how digital humanities can respond to cultural emergencies in the age of artificial intelligence, not only by archiving the past, but by creating accessible, collaborative spaces that preserve and expand living traditions.

Colleen Nugent McLean is the Coordinator of the CDS and the Project Manager for DHNow, a weekly academic newsletter highlighting works in digital humanities that are missed by traditional academic publishing. As CDS Coordinator, she led a crowdsourcing initiative to create documentation for projects to use platforms like Zooniverse. The relaunched DHNow is an important community resource in the age of fragmented social media. For this panel, she will speak on the importance of connection and community across projects, methods for strengthening these connections, and using crowdsourcing in times of insecurity. Caitlin Pollock is the Associate Director of the Digital Scholarship Group (DSG) and codirects the Boston Research Center, a Mellon-funded initiative to tell the community histories of underrepresented neighborhoods in Boston. As the Associate Director, she also works to further build relationships with the Center's resident projects, all of which the DSG provides technical support. For this panel, she will speak on connecting technical resources and systems to projects, building connections and partnerships in the face of limited resources and capacities, and creating welcoming events at the Library for Boston community members. Shireen Zaineb is the Digital Projects Archivist for the Humanities Center, where she provides archival support to the Humanities Center’s Signature Projects and emerging digital projects across the university. Working in a role that necessitates movement between teams and Centers, she will discuss the importance of consistent documentation, resource sharing, and event planning to cultivate knowledge transmission across projects. Her collaborations include the cultivation of a data resilience plan with the Digital Archive of Indigenous Language Persistence (DAILP) Project Coordinator Cara Hullings. Cara Hullings is the Project Coordinator for the Digital Archive of Indigenous Language Persistence, a grant-funded, community-based digital archive created to support indigenous peoples’ knowledge, interpretations, and recuperations as it relates to language. There, she is the point person for administration, project management, and communications. Cara will focus on the ways in which DAILP embeds communication into its open-source software and data management to facilitate the connections and cross-collaboration needed to remain resilient.

The digital humanities – as a heterogeneous collection of methods, pedagogies, and disciplinary formations – is no stranger to debates over the approaches contained under its umbrella. Ethnographic approaches to the digital, however, are often conspicuously absent from these discussions (with isolated exceptions such as Clement 2016 and Hsu 2014). In this talk, I outline the history and methods of ethnographic approaches to the digital, focusing on digital ethnography, multimodal ethnography, and ethnographic science and technology studies. Focusing particularly on digital ethnography, which takes digitality not only as its object of study but also its method, I argue that the ongoing divide between ethnography and the digital humanities is the result of historical disciplinary boundaries more than any concrete methodological or philosophical differences. The sparse commentary that does exist on this divide typically attributes it to ethnography’s own predilection for studies about rather than with technology (e.g., Varis 2020), a perspective which fails to account for a wide range of historical and contemporary ethnographic research practices. Ethnographers have already been moving towards DH – from ethnographic studies of DH (Bailey 2015) to growing calls for ethnographers to embrace DH methods (Fraser & Linares 2022) – and it is time for digital humanists to, in turn, move towards ethnography. I draw on my experience conducting ethnographic research in virtual worlds and mixed reality environments, as well as teaching both DH and ethnographic methods to undergraduates, to demonstrate how ethnography can supplement not only our DH methodological toolbox but also help transform the way we teach DH in our classrooms. Following from this year’s conference theme, the ethnographic study in, of, and with digital cultural phenomena offers DH a set of robust methods for studying what digital ethnographer Celia Pearce terms “emergence” (2009). Decades of poststructuralist ethnographic research and theory have demonstrated that human interaction and experience are not predetermined outputs of cultural, social, and technological systems, but rather a set of flexible, contingent performative engagements between porous subjects. “Emergence,” in how the term is used in digital ethnography, offers a robust methodological-theoretical nexus for better understanding the always-contingent nature of human engagement with sociotechnical apparatuses. Of particular interest to DH is how the digital ethnographic study of emergence offers a reflexive approach to method and theory within and around digital technologies, fine-tuned through deep engagement with the insights of feminist, postcolonial, and abolitionist ethnography; as such, digital ethnography offers one possible answer to questions of reflexivity and care towards which much recent work in DH has turned (e.g., Fan 2018; Juhasz 2021; Nowviskie et al. 2019; Sano-Franchini 2015).

The Anti-Racist Digital Research Institute (ARDRI) at the University of Michigan (U-M) began in 2022 as a mini-grant program that supported early-stage digital scholarship projects that advanced anti-racism and social justice in the humanities, arts, and humanistic social sciences. Since then, ARDRI has evolved into a yearly summer institute and grant program built on a thriving research community. The summer institute brings together a cohort of six research projects committed to using ethical and reparative frameworks that counteract historical and ongoing harm, particularly focusing on communities disproportionately affected by historic and systemic racism and neglect. The Institute has supported a total of twenty four digital scholarship projects on a range of topics, including, a collective memory and community engagement project surrounding the Detroit River, an educational platform for supporting Black girls experiencing early puberty, a digital archive of the Palestinian diaspora, and more. With the proliferation of data exploitation, disinformation, and algorithmic violence, developing digital projects with care and justice integrated at every level is more critical than ever. The ARDRI Curriculum integrates the Anti-Racist Research Framework developed by Goings et al. with digital scholarship praxis and frameworks like the Data Ethics Canvas (Terrant et al. 2021), Consentful Tech Project (Lee & Toliver, 2017), and Data Feminism (D’Ignazio & Klein, 2020). Throughout the program, participants learn practices for protecting research participants and engage research communities as partners, rather than more traditional extractive research paradigms. A primary focus of this paper is the development of the workshop at the beginning of the summer Institute, which introduces participants to the frameworks while helping them develop a rough but tangible plan for their digital project. As participants learn the Digital Scholarship Project Lifecycle, they create an anatomical diagram of their project, outlining their desired outcomes, the deliverables they will create, and the communities who are involved in and affected by their research. The participants then connect those elements to map out each phase of the project and how they connect, and crucially, how they will work with their communities at stake to create their project. The next phase of the Anti-Racist Digital Research Institute turns its focus outward. After five years of supporting, teaching, and refining the methods of doing Anti-Racist and digital work, we seek to share what we have learned and engage with a broader audience. This presentation reflects the program from its emergence and its development over the years, and assesses its impact. We share successes, lessons learned, and resources to help digital humanities practitioners implement and expand on them. Our goal, in both this paper and the Institute itself, is to support digital humanists and community members so that oppressive and exploitative models may be eschewed in favor of a praxis that is both equitable and just.

This paper presents Sensing Revolutions, a digital humanities project that responds to political emergencies and times of crisis by rethinking how knowledge about protest is produced, represented, and shared. Drawing on visual, auditory, spatial, and ethnographic materials gathered during the 2019-2020 Lebanese protests, the project uses interactive ArcGIS StoryMaps and a WordPress-based platform to reframe protest as a multisensory and relational experience rather than a spectacle of violence or disruption. Sensing Revolutions functions simultaneously as open-access, peer-reviewed digital scholarship as well as a pedagogical resource, featuring over 200 photographs, 30 videos, audio recordings, field notes, and archival materials contextualized through research and teaching toolkits. Framed within the ACH 2026 theme of emergence/ia, this paper understands protest as both emergency—a response to urgent economic and political conditions—and emergence, a medium for generating forms of collective sensing, solidarity, and public pedagogy. We argue that conventional written narratives and academic formats are insufficient to convey the lived realities of protest. These realities are inherently embodied, multisensory, and situated in urban space. By contrast, digital methods allow for historically grounded and experiential narratives that integrate sound, image, movement, and mapping into coherent multimodal narratives. Here our work aligns with other initiatives in digital humanities (i.e. The Syrian Archive, Filming Revolution, Memorias de la Resistencia) that support critical inquiry and social justice scholarship in the region and beyond. The project draws on theories of activist art, urban studies, and critical digital humanities to challenge reductive media portrayals of protest as sites of chaos and violence. Instead, Sensing Revolutions foregrounds protest as a space of dialogue, embodied dissent, care, political education, and conflict. A core commitment of the project is multilingual access, with materials presented in Arabic and English through glossaries, translations, and layered explanations designed to support non-specialist audiences. We position this multilingual, multisensory design as a form of ethical digital humanities practice, which is attentive to accessibility and the politics of representation. In addition to creating an immersive educational experience, the project will contribute new, interactive pedagogical tools for classrooms at the undergraduate levels. Through lesson plans, multimedia assignments, and contextual glossaries, Sensing Revolutions equips educators with resources for teaching about revolutions, protests, and socio-political struggles in innovative and interactive ways. By centering voices from the Global South—regions often marginalized in US educational contexts—the project foregrounds complexity, nuance, and lived experience, offering a more inclusive and relational understanding of modern history. This paper ultimately argues that Sensing Revolutions offers a transferable model for digital humanities work across different geographies, particularly in contexts shaped by protest, environmental crisis, economic instability, migration, and political precarity. By foregrounding sensory experience, public digital humanities, and multimedia storytelling, the project contributes to broader conversations about how digital humanists respond to emergencies while cultivating emergent forms of knowledge, solidarity, and care.

In October 2019, The New Arab reported on the launch of the “world’s first” artificial intelligence fatwa service in Dubai. Virtual Ifta’ was described as able to deliver a fatwa – a religious jurisprudential ruling, issued by a qualified Islamic religious legal expert in response to a specific question – in answer to approximately 200 different questions related to prayer, drawing from its database of 130,000 fatwas. Intended to reduce the workload of the UAE’s Islamic Affairs Department, the program was discontinued in 2021 as the Emirati government shifted towards business-oriented AI applications. The following year, the launch of ChatGPT ushered in the start of a new era of experiments in generative AI-based fatwa applications. This paper examines the emergence and impact of generative artificial intelligence (GenAI) fatwa generators, also known as fatwa engines. It contributes to the arena of digital humanities due to its use of humanistic research into digital objects and culture, and to the field of digital Islam by offering a survey and analysis of these fatwa-generating applications. It uses small-scale data mining as well as visual and textual content analysis to understand the characteristics of these fatwa generators, and draws upon relevant literature in both fields in its analysis. It takes a case study approach, focusing on applications like AI Mufti, a ChatGPT-powered “ethical fatwa engine” that promises to answer “all your Islamic legal and ethical questions” based on Qur’an, hadith, fiqh, and the consensus of the four Sunni madhhabs. Despite the references to muftis and fatwa engines, the creator states: “This assistant doesn’t replace a qualified mufti”, illustrating the careful line drawn between AI fatwa generation and the issuance of fatwas by qualified, human muftis. Similarly, the Hanafi Fatwa Document Reply Generator suggests with its name that it offers fatwas; it offers fatwa-style responses to questions asked, in accordance with the Hanafi madhhab – but with the disclaimer that they are not actual fatwas. Yet in both cases, the application’s name suggests the exact opposite: that they are authentic and credible sources of Islamic religious jurisprudence. These applications and their framing suggest the appeal of making fatwas more accessible to ordinary Muslims and allowing them to request infinite rounds of clarification – or new questions – of a tireless digital cleric. However, the paper argues, the disclaimers offered by each application reveal a reticence to fully blur the historical distinctions between fatwas as authoritative, actionable jurisprudence and all other religious advice, no matter how well intended or grounded in scripture. Instead, while their titles promise fatwas, their descriptions frame them as educational, informational tools, or even research assistants. This paper offers two contributions. First, it adds to our understanding of digital Islam and its impact on contemporary Muslim beliefs and practices – including how these can become interwoven with state and corporate interests. Second, it offers a replicable model for analyzing genAI-based applications offering religious guidance historically provided by a religious cleric or other trained professional, thus also supporting work in religion-focused digital humanities.

Emergence/ia asks digital humanists to create and collaborate through moments of urgency—balancing “emergence” (new methods and connections) with “emergency” (care, response, and transformation). Our project directly addresses the themes of care, response, and transformation in digital humanities. We "care" for the cultural heritage of East Germany by digitizing and analyzing Mosaik, a primary educational and entertainment medium for GDR youth. Our "response" involves developing specialized digital tools for formats that deviate from standard comic structures, effectively "transforming" traditional archival research into a data-driven, multimodal inquiry.     Our proposal presents a novel multimodal analysis pipeline for the Mosaik series, positioning it within the context of East German Bildgeschichte (picture stories). By integrating advanced panel segmentation and generative AI (GenAI) tools, we explore the socio-political nuances, censorship, and religious themes embedded in this unique medium.     We situate our work in current multimodal comics-understanding research, including the Pick-a-Panel framework (e.g., ComicsPAP [Vivoli et al.]), as a point of comparison, while arguing that Bildgeschichte demands new segmentation and sequencing priors before higher-level reasoning can be evaluated meaningfully.     To achieve that, we develop an end-to-end, semi-automated segmentation-to-interpretation pipeline for the scanned Mosaik EPUB sourced from online libraries or institutional digitizations. Our workflow begins with scanning Mosaik Comics from the resource to combine with the formal existing digitizations. With the corpus ready, we build the metadata table in capture of all series of Mosaik and their relational information. The segmentation pipeline is built around the semi-automatic idea of using detection-and layout-aware strategies (YOLOv12-based detection plus gutter/pointer cues) with the help of human annotation to tailor Bildgeschichte’s atypical composition.     The resulting structured outputs—panel crops, speech-text regions, and OCR text—feed a lightweight dataset schema that supports downstream multimodal generative-AI–assisted analysis. Specifically, we experiment with “reverse prompting” to elicit interpretable intermediate descriptions (scene/object categories, character presence and agency, and interaction types) grounded in both panel imagery and extracted text, and we log the model's uncertainty and failure to keep interpretation accountable. Substantively, we use this infrastructure to build the panel table that includes necessary information to proceed with the textual analysis and story-arc analysis for the literature review.     In the literature review, our practice will try to unearth critical questions regarding censorship and political awareness in East Germany. The pipeline tracks geographical and chronological travel within the narratives, revealing how the GDR portrayed the West and historical eras. Furthermore, we analyze the emergence of religious and spiritual themes within a theoretically secular socialist state.     Methodologically, we position our work as a “care practice” for digital cultural heritage: making historically contingent visual narratives computationally legible without flattening their sociopolitical context under the enhancement of emerging technologies.     The significance and value of this work lie in that it provides the digital humanities field with a reproducible computational framework for studying culturally specific multimodal media, reproduces the literary scholarship with new tools and methodology, and offers new insights into the sociopolitical implications of East German educational media.

This paper presents the process and results of the first-ever Zooniverse framework for ethical integration of Artificial Intelligence and Machine Learning (AI/ML) into online crowdsourcing. The Zooniverse (https://www.zooniverse.org) engages nearly 3 million volunteers worldwide who have contributed to 450+ participatory research projects, each led by a different research team. The field of crowdsourced research is mature, with clearly-articulated best practices for equitable approaches which prioritize volunteer experience at the same level as data production (Ridge et al. 2021). The rapid introduction of widespread AI/ML has the potential to disrupt current practice if not approached with caution and intention. Examples of AI/ML engagement can include: incorporating ML models for data pre- or post-processing; using volunteer-labeled data to train ML models; and “Human-in-the-Loop” processes which combine these approaches (Fortson et al. 2024). These methods are increasingly prevalent as the technology increases in sophistication and accessibility for researchers and the public alike. As the use of AI/ML has become more prevalent—now in ⅓ of Zooniverse projects—it has sparked a range of reactions within the volunteer community, reflecting broader societal discourse. On project discussion forums, Zooniverse participants have surfaced concerns on issues like ownership, agency, data quality, and transparency. In spaces where public participation is based on trust, unchecked adoption of new technologies can potentially lead to a breakdown of existing social contracts and other negative impacts (e.g. to people, to research data, to carefully-developed methodologies). In response, Zooniverse launched an initiative to develop recommendations for running AI/ML-engaged projects on the platform. This included a virtual workshop series that brought together a range of stakeholders, including platform leadership, project teams, volunteers, ethicists, and experts in AI/ML, participatory research and related topics. Workshops covered transparency and communication best practices to support researchers in effective public communication; foundational elements of ethical research; factors that influence benefits and risks including disciplines, task types, and varied stakeholder needs; and downstream data protection in an age of web scraping and generative AI. To ensure that our large volunteer community was broadly represented, all workshops were preceded by global email surveys to gather additional input from Zooniverse volunteers about their awareness of when and how AI/ML is used in Zooniverse projects; potential benefits and risks of their incorporation; and preferences and recommendations for how to proceed. Here we share the resulting Zooniverse framework for AI/ML-engaged online crowdsourcing, which will be incorporated into platform policy and process and will continue to be refined through a beta testing and iterative improvement process. We hope the usefulness of this work extends to other platforms and researchers working at the intersection of public participation and AI/ML, and that we can demonstrate the importance of a careful response to rapidly-shifting technological change. This work contributes to ongoing discourse in the DH community as well as the themes of the 2026 ACH conference by exposing the challenges and opportunities of creating frameworks for the widespread adoption (or avoidance) of a new technology in public-engagement driven research spaces which prioritize transparency, communication, and equitable participation.

This Beautiful Sisterhood of Books, a digital humanities project, recreates the Woman’s Literary Department from the New Orleans 1884 World’s Fair and “recovers the exclusion” of Black women and their work from the Woman’s Department exhibit.[1] The project encourages scholars, educators, students, and the public to explore and interpret the collections and resources. The site offers a digital archive, an undergraduate curriculum, and exhibits relating to the 1884 Woman’s Department and late-19th century Black women writers. This paper will provide a short history of the project, visualize the fair data, and illustrate our process for building and engaging the public and students through digital history practices. The paper covers keyword/metadata creation, database structure, storage, types of exhibits (archives, interactive essays, timelines, story maps), crowd-sourcing techniques, and curriculum integrations. We argue that projects like the Beautiful Sisterhood crucially link public engagement with intersectional histories and represent an expanding process of knowledge production by and for an interdisciplinary community.[2] DH projects, like the Beautiful Sisterhood, serve as a critical response to the on-going emergency that humanities educators and the humanities fields currently face, especially when presenting. U.S. history topics about race and gender using established historical practices in digital spaces. The original Woman’s Literary Department was created for the 1884 New Orleans World’s Fair. It comprised more than 1,400 items (books, journals, newspapers, and sheet music) donated by publishers and collected by “lady delegates” from around the US. Black women were not represented, despite having proposed to exhibit their works alongside those by white women. The Woman’s Literary Department’s director, Maud Howe, declared herself determined “to keep this beautiful sisterhood of books together” as a permanent library. Within months, however, the physical collection was scattered and lost.[3] More than 130 years after the fair, the Beautiful Sisterhood collection has reemerged through digital archival processes using Maud’s shelf list. To confront Julia Ward Howe’s failure to integrate the collection, we focused on recovering the presence and absence of Black women. Simply adding Black women’s publications into the database was unsatisfactory. We wished neither to erase the historical fact of the 1884 collection’s Jim Crow instantiation, nor to downplay the racism of the white women organizers and exhibitors. It seemed crucial to preserve, in some form, the fairgoers’ original perspectives in the post-Reconstruction South. Our interactive DH exhibit, “Black Women and the 1884 World’s Fair,” links to the database but also stands on its own as what Kim Gallon terms an “alternative construction.”[4] Some may view the creation of a digital collection as moving archival items from shelves to servers for accessibility and preservation. Yet, this process of presenting historical events through documents and digital tools creates new digital methods for assembling and engaging with archival materials and new ways of knowing historical subjects that situate old and new knowledge together.[5] [1] Adams and Howard, https://thisbeautifulsisterhood.org/ [2] Haraway, “Situated Knowledges: 581–82. [3] Weimann, The Fair Women, 391; Pfeffer, Southern Ladies, 79. [4] Gallon, “Making a Case,” 44. [5] Haraway, “Situated Knowledges: 581–82.

Since 2017, the Global Medieval Sourcebook (https://sourcebook.stanford.edu/) has served as a venue for disseminating English translations of medieval-era texts from around the world. Developing proficiency in reading texts in historic languages is an important yet time-consuming aspect of the medieval studies curriculum. Translating less well-known texts – individually or in a group – is a common way to practice. Scholars translate texts for classes, in order to give students without the linguistic background access to less-common primary source materials. Translation is undervalued, and there are few places to publish these pieces in an institutionally-legible way. As a result, most never see a wider audience. The Global Medieval Sourcebook was designed to offer wider visibility to English-translated works spanning six centuries of literary production across Europe, North Africa, the Americas, and Asia that convey the diversity and vibrancy of medieval culture. A 2018 grant from the National Endowment for the Humanities supported the development of a Drupal 7-based platform, with custom code for displaying TEI-encoded versions of the original text and translation. The sustainability of the project was jeopardized by the departure of the original developer soon after the site was created, along with the complete technical overhaul of Drupal 8, which came with more intensive server requirements and would have necessitated rewriting all the custom code from scratch. Responding to these changes, the GMS team reimagined the project as a one-time anthology in the form of a static Jekyll site with meticulously designed PDFs oriented towards printing, and adapting the Digital Humanities Literacy Guidebook (https://cmu-lib.github.io/dhlg/) into a form suitable for translations. The last 6 years of the project’s development, culminating in the spring 2026 release of a new static site built in 11ty with 100 texts, offers several lessons for scholars looking to undertake a similar project in the current funding and institutional context for the humanities. Among them is the significant need – and cost – for expert human labor in preparing and proofing (especially non-English original) texts for publication, in each modality (print and digital). AI tools offer little to no meaningful support in this process, but can significantly reduce the technical burden in setting up a static site (see Dombrowski 2022) or migrating from one static platform to another. Prioritizing printable and web-displayable outputs rather than directing resources to TEI markup when none of the project participants use TEI for their research was also an important decision point. The minimal system we have built, which responds to the provocations offered by Roopika Risam’s DH 2025 keynote (Risam 2025), ironically has allowed us to consider expanding the project from a one-time, completed anthology to a flexible, ongoing publication that can provide DOIs for translations. Our talk will reflect on these lessons with an eye towards their applicability to other projects that preserve and distribute texts in the context of limited funding and infrastructure. References Dombrowski, Quinn. “Minimizing Computing Maximizes Labor”. Risam, Roopika. “Digital Humanities for a World Unmade”. Digital Humanities 2025 keynote. Velloso-Lyons, Mae “Global Medieval Sourcebook.” Seminar: A Journal of Germanic Studies.

Historical directories—ranging from city registers to specialized professional rosters—represent a pervasive yet fundamentally underutilized archival resource. These documents provide remarkably granular, individual-level data concerning historical demographics, professional dynamics, and business locations that are essential for history and social science research. However, the availability of these records in printed form has constrained researchers to labor-intensive manual data entry or limited sampling methods, which often prioritize major urban centers while leaving rural and marginalized populations underrepresented. This paper details a work-in-progress project that addresses the "emergency" of archival inaccessibility by fostering the "emergence" of a scalable, automated pipeline for directory digitization. The proposed methodology utilizes advanced artificial intelligence and machine learning frameworks to move beyond mere optical character recognition (OCR) toward the creation of structured relational databases. The pipeline consists of four distinct phases: (1) high-definition text and layout extraction from scanned images; (2) the training of custom extraction models to distinguish directory content from peripheral noise such as advertisements, headers, and footers; (3) the programmatic segmentation of raw text into discrete, multi-line entries; and (4) the parsing of those entries into defined data fields based on syntax and punctuation. While contemporary enterprise AI tools are often optimized for standard business forms and invoices, historical directories present a unique methodological challenge due to their dense shorthand, irregular column structures, and inconsistent archival typography. Central to this project is the role of the academic library as a site for digital scholarship and pedagogical innovation. The project provides a robust framework for experiential learning, where students with strong computational backgrounds apply advanced tools to solve real-world archival problems. Through this mentored research, students gain expertise in managing data integrity, evaluating error propagation, and refining "human-in-the-loop" validation processes necessary for high-accuracy historical digitization. The ultimate output—a generalizable pipeline for converting directory PDFs into structured data—enables scholars to conduct sophisticated statistical inquiries and track professional trajectories across vast temporal and geographic scales. This research furthers the strategic library objectives of preserving cultural heritage while empowering a new generation of researchers with the skills to navigate and transform the digital archival landscape.

Political tensions in the United States are rewriting, erasing and threatening the fields of Latino studies and the documentation of ethnic histories. Cultural memory is under threat not only from natural disaster, but also from political erasure, institutional neglect and growing precarity. Archival scholars such as Mario H. Ramírez demonstrate that archival silences are produced through racialized institutional practices that marginalize Latino histories. “Emergencia,” thus, signals not only rapid response to preserve precarious materials, but also the structural conditions that make safeguarding cultural memory an urgent and ongoing practice. This panel examines how community archives and US Latino DH projects build infrastructures that move beyond digitization and metadata toward embodied relationality. As Baeza Ventura et al. argue, “Latinx Digital Humanities centers Latinx lives, community, intellectual production, scholarship and archival collections” and “engages the ethical protocols developed by ethnic studies and feminist practitioners.” Preservation becomes an ethical and relational act that allows marginalized voices to emerge. Digital work functions as long-term cultural stewardship, placing US Latino archival histories into conversation with canonical narratives and expanding how we conceptualize US American history and culture. Safeguarding US Latino cultural memory is not supplemental to digital humanities, but a critical intervention that reshapes how the field understands preservation, infrastructure and public knowledge. "Shedding, Emergence and Nepantla: Transformative Practices in US Latino DH" Drawing on Gloria E. Anzaldúa’s concept of nepantla, this presentation theorizes preservation as continual adaptation. In this in-between space, practitioners must shed inherited archival frameworks and emerge with new strategies responsive to political and institutional pressures. Cultural heritage workers safeguarding Latino histories must remain agile, reimagining how archives are described, contextualized and activated for public engagement. "(Digital) Methodology of the Oppressed" Grounded in decolonial and women-of-color feminist thought, this presentation introduces a “digital methodology of the oppressed,” drawing on María E. Cotera and others. Centering community-based knowledge, embodiment and affect, this framework resists colonial narratives and moves digital humanities beyond textual analysis toward connection, witnessing and remembrance. "Community Work as Infrastructure" This presentation traces how the US Latino Digital Humanities Center (USLDH) emerged through outreach, collaboration and crowd-sourced lists of digital projects. Infrastructure-building became community-building, or a relational network that sustains preservation beyond and as emergency response. "Public-Facing Preservation through Ethics, Access and Accountability" This presentation examines how community archives, bilingual projects and outreach expand access to US Latino archives beyond academia. By designing materials in both Spanish and English, preservation serves the communities whose histories are being safeguarded. Public engagement is a core preservation strategy grounded in reciprocity and accountability. Anzaldúa, Gloria. Borderlands/La Frontera. 2nd ed., Aunt Lute Books, 1999. Baeza Ventura, Gabriela, et al. “A US Latinx Digital Humanities Manifesto.” Debates in the Digital Humanities 2023, edited by Matthew K. Gold and Lauren F. Klein, University of Minnesota Press, 2023. Caswell, Michelle. Urgent Archives: Enacting Liberatory Memory Work. Routledge, 2021. Cotera, María Eugenia, et al. “The Archive as Encuentro.” Chicana/Latina Studies, vol. 15, no. 2, 2016. Ramírez, Mario H. “Being Assumed Not to Be.” The American Archivist, vol. 78, no. 2, 2015, pp. 339–356.

While DH sustainability challenges are well documented and studies of approaches to sustainability are increasing (e.g., Goddard, 2023), empirical clues to the full scale of the problem are sparse. The Project Endings survey found that “only about 10% [of projects] were archived in a stable, long-term environment” (Goddard, 2023). A survey of articles in Digital Humanities Quarterly found that “31% of citations containing links…no longer work correctly” (Coble & Karlin, 2023). Davis (2015) found that 45% of digital humanities projects were no longer accessible online 10 years later. These numbers are troublesome and suggest the need for a broader empirical study of the scale, scope, and impact of unsustainability of digital cultural evidence.   The DisappearingDH project began with a question adapted from the Internet Archive report on Vanishing Culture (Messarra et al., 2024): What is the extent and cultural significance of DH resources that have disappeared or are at-risk? In stage one we gathered data about digital cultural projects from the databases of two U.S. federal funding agencies, the Institute of Museum and Library Services and the National Endowment for the Humanities. While our data collection will expand beyond particular funders and national contexts, the urgency of existential political threats to both agencies (and their databases) in 2025 spurred our initial focus. This presentation reports on a growing dataset of 440+ IMLS- and NEH-funded DH projects and their outcomes, which supports longitudinal study of the sustainability of digital cultural resources. We will detail methods of data collection, including sampling, inclusion criteria , and manual qualitative coding of project outcomes and their status. While we set out to characterize the “half-life” of DH resources, our close analysis of projects and their trajectories has challenged our thinking toward a more interpretive set of questions: reckoning with conceptual variations among completeness and resilience (Edmond & Morselli, 2020; Holmes et al., 2023); data journeys (Leonelli & Tempini, 2020); maintenance (Mattern, 2018); and impact stories (Marsh et al., 2016). Coble, Z., & Karlin, J. (2023). Reference Rot in the Digital Humanities Literature: An Analysis of Citations Containing Website Links in DHQ. Digital Humanities Quarterly, 017(1). https://dhq-static.digitalhumanities.org/pdf/000662.pdf Davis, R. C. (2015). Taking Care of Digital Efforts: A Multiplanar View of Project Afterlives. Publications and Research. https://academicworks.cuny.edu/jj_pubs/231 Edmond, J., & Morselli, F. (2020). Sustainability of digital humanities projects as a publication and documentation challenge. Journal of Documentation, 76(5), 1019–1031. https://doi.org/10.1108/JD-12-2019-0232 Goddard, L. (2023). What’s Left When It’s Over: Libraries and Digital Humanities Project Preservation. IDEAH, 3(5). https://doi.org/10.21428/f1f23564.00188fc5 Holmes, M., Jenstad, J., & Huculak, J. M. (2023). Introduction to Special Issue: Project Resiliency in the Digital Humanities. Digital Humanities Quarterly, 017(1). Marsh, D. E., Punzalan, R. L., Leopold, R., Butler, B., & Petrozzi, M. (2016). Stories of impact: The role of narrative in understanding the value and impact of digital collections. Archival Science, 16(4), 327–372. https://doi.org/10.1007/s10502-015-9253-5 Mattern, S. (2018). Maintenance and Care. Places Journal. https://doi.org/10.22269/181120 Messarra, L., Freeland, C., & Ziskina, J. (2024). Vanishing Culture: A Report on our Fragile Cultural Record. https://blog.archive.org/wp-content/uploads/2024/10/Vanishing-Culture-2024.pdf

Minimal computing has reshaped digital humanities by foregrounding sustainability, labor equity, and infrastructural restraint (Gil 2015; Sayers 2016; Gil and Ortega 2016; Risam 2019). At the same time, this tool-centered articulation has encouraged a widespread assumption that relational databases are incompatible with minimal computing because they are too complex, too dependent on specialist labor, too fragile for long-term preservation, and too resource-intensive for low-bandwidth contexts. As Risam and Gil (2022) argue, minimal computing is better understood as a values-driven design philosophy grounded in necessity and sufficiency rather than particular tools. Their rubric — What do we need? What do we have? What must we prioritize? What are we willing to give up? — raises an underexplored possibility: what if a project needs a relational database? Drawing on our experience building the Pan-African Data Project, a digital mapping project documenting Pan-Africanism from the late nineteenth century to the present, this paper proposes “hybrid minimal computing” as a design approach that sequences relational systems over time. Pan-Africanism survives in fragile and dispersed archival traces shaped by colonialism, state surveillance, and uneven preservation. Representing its transnational histories requires maintaining relationships among individuals, organizations, events, and places across space and time. A purely page-based model would reproduce the fragmentation of the archive itself. Yet a continuously server-dependent database introduces long-term risks — hosting costs, maintenance burdens, institutional dependency, and infrastructural precarity — that conflict with commitments to sustainability and accessibility, particularly for users in low-bandwidth environments. Hybrid minimal computing addresses this tension by separating relational modeling from public deployment. During active research, the project uses a conventional relational database to preserve normalized records and explicit relationships. At publication, however, maps, timelines, tables, and narrative materials generated from this relational core are compiled and delivered as a fully static site. Relational complexity is retained where necessary and released where it would otherwise impose ongoing infrastructural risk. This approach reframes minimal computing as infrastructural sequencing rather than technological austerity. It treats constraint as a design resource and responds to contemporary emergencies of funding volatility, environmental cost, and global inequity in digital access. By formalizing when and why relational systems are used — and when they are intentionally abandoned — hybrid minimal computing offers a generalizable, open-source model for digital humanities projects that require relational rigor without permanent server dependence. References Gil, Alex. 2015. “The User, the Learner and the Machines We Make.” Minimal Computing Working Group, May 21, 2015. https://go-dh.github.io/mincomp/thoughts/2015/05/21/user-vs-learner/. Gil, Alex, and Élika Ortega. 2016. “Global Outlooks in Digital Humanities: Multilingual Practices and Minimal Computing.” In Doing Digital Humanities: Practice, Training, Research, edited by Constance Crompton, Richard J. Lane, and Ray Siemens, 58–70. Routledge. Risam, Roopika. 2019. New Digital Worlds: Postcolonial Digital Humanities in Theory, Praxis, and Pedagogy. Northwestern University Press. Risam, Roopika, and Alex Gil. 2022. “Introduction: The Questions of Minimal Computing.” Digital Humanities Quarterly 16 (2). https://dhq-static.digitalhumanities.org/pdf/000646.pdf. Sayers, Jentery. 2016. “Minimal Definitions.” Minimal Computing Working Group, October 2, 2016. https://go-dh.github.io/mincomp/thoughts/2016/10/02/minimal-definitions/.

The rapid emergence of Large Language Models presents an emergency for humanistic education. LLMs excel at summarizing, ranking, and generating plausible text — but these very capacities threaten to make close reading obsolete. When students and scholars rely on AI-generated summaries rather than direct engagement with texts, critical thinking erodes. The ability to form independent questions, gather evidence, and construct interpretive arguments is displaced by algorithmic selection that chooses for the reader rather than supporting her. Attempts to move beyond the close/distant binary have diversified what computation does with literature — Piper's alternative reading modes (2018), Underwood's conciliatory "new scale of description" (2019) — and a smaller tradition has applied computational methods to individual texts (Eve 2019; Sá Pereira 2019). Yet even these single-text studies use computation to observe formal features. None uses computation to construct a new text as a way of experimentally testing how literary form produces meaning. Our approach, structural experimentation, extends the practice of "deformance" — the deliberate deformation of a text to release its latent formal logic — theorized by McGann and Samuels (2001) and advocated by Ramsay (2011) — by extracting the structures operating at the semantic and topical levels and applying them to new content, testing their meaning-producing operations in an alien domain. Novels have proven to be an excellent benchmark for assessing AI's contextual processing of extended texts (Wang 2024; Bonomo 2025). Current LLMs still face significant limits when processing long narratives: models with context windows larger than 200K tokens struggle when provided context exceeds 100K tokens, and their performance on meaning and relevance tests fails to reach 30% (Wang 2024). Operating on computationally decomposed formal templates extracted from discrete textual units, our approach sidesteps the context-window problem entirely while enabling a genuinely new mode of AI-assisted criticism. We demonstrate this through a technique of "intertextual mapping" applied to Alessandro Manzoni's The Betrothed (~420K tokens). The novel's opening chapter founds a shared, future-oriented historic space through specific formal operations (Raimondi 1975, 1987). We computationally decompose its formal architecture — syntactic hierarchy, perspectival sequence, and grammar of autonomous agency — into a template specifying the preservable structural properties of Manzoni's prose. This template is then applied to a different foundational scenario: the transformation of a contemporary Italian urban neighborhood. The AI performs formal parsing, constraint monitoring, and alternative generation; the human scholar provides philological judgment and evaluates the results. By detaching Manzoni's formal architecture from its original content and applying it to politically charged contemporary material, the experiment is designed to make visible formal operations — such as the naturalization of process, perspectival authority, and the grammatical exclusion of human agents — that remain invisible when form and content are fused in the original. We propose that the current emergency contains within it an emergence: AI tools that support the relationship between formal attention and critical thought. Realizing this requires design choices that prioritize structural analysis over summary, experimentation over observation, and the scholar's interpretive agency over algorithmic selection.

Digital collections of texts by minoritized writers can do many things, from providing the bare bones of access to otherwise inaccessible materials to helping to document and validate marginalized communities as they represent themselves to the broader world. With buy-in and engagement from members of the affected communities themselves, they can also provide cultural context and a pedagogical framework that can make such materials legible to others. Here, I explore the advent of a recent Scalar project, Adivasi Writers: An Introduction to India’s Indigenous Literature. Adivasis are South Asian indigenous communities, subject to a long history of marginalization and displacement going back to the colonial era and continuing in the present. There are more than 100 million Adivasi people in India alone, located throughout the country, with particular concentrations in central India as well as in the northeast region. For generations, Adivasis were written about rather than subjects of their own story. Adivasi Writers aims to help change that by centering Adivasi voices directly in a decolonial framework, and making their writings accessible to a broad readership. Our site centers the voices of Adivasi writers like Hansda Sowvendra Shekhar, Vandana Tete, Nirmala Putul, Anuj Lugun, and Gladson Dungdung, while drawing ethnographic scholarship on Adivasi literatures and cultures by scholars like Ashok Kumar Sen, G.N. Devy, Daniel J. Rycroft, and Sangeeta Dasgupta. Aside from simply compiling a list of writers grouped by preferred community (surprisingly, no such collection appeared to exist prior to our project), here I describe how Adivasi Writers charts the growth of a concept for emergent identity formation, Adivasiyat (or “Adivasi-ness”). For many writers, this concept supercedes a previous focus on individual communities or tribes, and works as a strategic pan-ethnic framework to support Adivasi civil rights, Adivasi claims over land, and Adivasi ecology. Our site uses thematic tags as well as pedagogically-oriented contextual essays to guide readers outside of these communities into Adivasi writing as a scene of emergence.

The L.A. River is often called a “joke,” given how thoroughly it defies the paradigms of a river set by Eastern U.S. or European standards. Rather than meander, its fifty-one-mile path is nearly straight, flowing down the middle of gaping concrete slopes. A desert river, with highly irregular flow, it’s often little more than a trickle, adding an element of absurdity to its infrastructural footprint. However, like other marginal rivers, the importance of the L.A. River is recently being re-examined, given that Los Angeles is increasingly prone to wildfires and severe flooding, and revitalization of sections of the river and its tributaries could create a safer and more sustainable city. In this paper, I examine how the river was initially turned into a joke as part of a rhetorical campaign to tap the Owens River (260 miles from L.A.) for the L.A. Aqueduct, during the period of 1900-1907, as evidenced in the pages of The Los Angeles Times. In the process, I explore the potential of a closely supervised OCR-correction pipeline using GPT4, in which I randomly check results against the newspaper scans. I also build on other work using BERTopic to examine newspaper datasets (such as Murugaraj et al, 2025) by adding in spaCy modeling (as explored by Kane, 2023 and others), allowing me to gain more granular insights from the data, such as the specific rhetorical modes employed. Datasets were gathered from 100-token snippets around mentions of the L.A. River in L.A. Times articles from Proquest’s TDM Studio, under their derived data terms of use. Ultimately, I found that the “joke” of the L.A. River cohered during this period, during which various rhetorical strategies were employed in order to belittle the L.A. River in favor of promoting the Owens River. These strategies included frequent sarcasm, appeals to scientific authority, and personification of the L.A. River that depicted it as undeserving of the people of Los Angeles. This project connects to my broader work exploring how urban rivers are fertile places in which to examine how a city constructs its own identity and its relationship to nature. I connect the analysis of this period to later newspaper datasets, including those from key flooding years, a period in which the river was promoted as a highway, and the modern-day, in which a revitalization effort is underway. Overall, this work examines how the way urban waters are talked about ends up determining their treatment.

Digital financial systems make split-second decisions about people’s money, and app communication; the automated messages users receive via in-app interfaces, email, SMS, and push notifications, becomes the only explanation users get. For those living paycheck to paycheck, these moments can escalate from technical glitch to genuine crisis like rent payments blocked, grocery purchases declined, emergency transfers delayed, and so on. This presentation examines how app communication operates as digital rhetoric during transaction failures, analyzing screenshots and message examples collected from fintech platforms across the Americas, Africa, and Europe. Using anonymized examples, this work demonstrates how language choices across channels shape user agency during financial disruptions. Specifically, the analysis examines: pronoun use (“we blocked this” versus “your transaction was declined”), tone (apologetic versus authoritative), sequencing (explanation before or after the denial), explanation depth (vague “suspicious activity” versus specific triggering factors), and action pathways (visible appeals and help versus dead-end messages). Rather than focusing on backend algorithmic models, the analysis centers app communication as the primary site where automated decisions become clear (or remain ambiguous) to users. Framed by ACH 2026’s theme of emergence/ia as both emergency and emergence, this study positions transaction failures as constituting immediate crisis for users while automated or AI-assisted messaging simultaneously emerge as everyday decision-makers in digital financial systems. The analysis asks: When automated systems make consequential decisions about people’s money, what role does language play in preserving or diminishing user agency and understanding? App communication is positioned as a form of humanistic knowledge-making that bridges computational systems and human experience by situating it within existing scholarly works in digital rhetoric (Rose & Tenenberg, 2015), critical algorithm and platform studies (Benjamin, 2019; O’Neil, 2016; Pasquale, 2015), and design justice (Costanza-Chock, 2020). This presentation calls on conference participants, UX practitioners and other researchers to recognize app communication as a critical site where language shapes agency and usability. Citations Benjamin, R. (2019). Race After Technology: Abolitionist Tools for the New Jim Code. Polity Press. Costanza-Chock, S. (2020). Design Justice: Community-Led Practices to Build the Worlds We Need. MIT Press. O'Neil, C. (2016). Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. Crown. Pasquale, F. (2015). The Black Box Society: The Secret Algorithms That Control Money and Information. Harvard University Press. Rose, E. J., & Tenenberg, J. (2015). Arguing about design: A taxonomy of rhetorical strategies deployed by user experience practitioners. University of Washington.

Can there exist a human right to remain unrecognizable within public space, and might such a right be exercised de facto by rendering oneself more visible to the human eye? If not, does the non-consensual application of facial recognition technologies constitute a direct infringement of the right to privacy? And if privacy is understood to include a sphere of biometric autonomy, could individuals legally and effectively resist their algorithmic recognition? This article analyzes CV Dazzle,[1] an avant-garde, anti-surveillance aesthetic as a crucial form of legal performance that exposes fundamental limitations in contemporary surveillance jurisprudence. The paper argues that the practice's significance is located in its performative failure, situating CV Dazzle as a semiotic signifier,[2] which serves as a demonstration of the breakdown of individual biometric autonomy and as a signifier to opposition to surveillance, to the algorithmic gaze and to facial recognition processing. By attempting to exercise the "right to hide" through visible dis-consent,[3] CV Dazzle's legal performance will be discussed theoretically and doctrinally. Theoretically, through Brighenti's notion of the vision,[4] as well as doctrinally, as resistance to the sufficiency in current surveillance law, demonstrates via the lens of the European Court of Human Rights' (ECtHR) privacy framework, and particularly the recent Glukhin v. Russia decision.[5] Specifically, it offers a normative rebuttal to the Court's reliance on the traditional "reasonable expectation of privacy" test. The analysis situates these aesthetic resistances within the context of rising technological authoritarianism,[6] particularly in populist regimes, where automated identification is strategically weaponized against political dissent and democratic rights. The paper concludes that CV Dazzle compels a necessary expansion of the legal imagination, demanding that the law move beyond individualized remedies to recognize the right to anonymity as a collective, democratic prerequisite for public life in the algorithmic age. [1] See the project description in Harvey's site: https://adam.harvey.studio/cvdazzle/. [2] Baal Delupi, Concealment of the Face and New Physiognomies, 19 Chinese Semiotic Stud. 511, 519 (2023). [3] Torin Monahan, The Right to Hide? Anti-Surveillance Camouflage and the Aestheticization of Resistance, 12 Comm. & Critical/Cultural Stud. 159, 160 (2015). [4] Andrea Brighenti, Visibility: A Category for the Social Sciences: A Category for the Social Sciences Current Sociology, 55(3), 323, 324 (2007). [5] Glukhin v. Russia, App. No. 11519/20, ¶ 82 (Eur. Ct. H.R. 2023) (Third Section). [6] Çisem Gündüz Arabaci, The Contestation of the Sensible under Populist Authoritarianism: Between Silencing and Seeking Democratic Rights, Front. Polit. Sci. 7, 6 (2025).

Periods of political, social, and national emergency generate intense demands for response, care, and collective meaning-making. In networked environments, these responses are frequently ritualized: publics repeatedly circulate familiar texts, phrases, and cultural forms in order to manage shock, signal ethical alignment, and stabilize affect. While digital humanities has developed robust tools for tracking circulation, virality, and engagement, such practices are often analyzed as informational or discursive flows rather than as ritual processes shaped by platform capitalism, affective governance, and emergency temporality. This paper argues for the need for a distinct methodological orientation—ritual analytics—for studying cultural circulation during crises. Where my related work examines specific poetic case studies in depth, this paper steps back to propose a generalizable methodological framework for studying ritualized circulation across platforms, genres, and crises. Ritual analytics integrates quantitative detection with interpretive and ethical frameworks drawn from media theory, digital anthropology, and Marxist cultural analysis. Rather than asking only what circulates or how much, ritual analytics asks how circulatory practices function socially: what forms of care they provide, what modes of action they preempt, and how they shape the conditions of public emergence. The paper synthesizes methods commonly used in cultural analytics—time-series analysis, event alignment, anomaly detection, and regression modeling—with interpretive approaches attentive to ritual, communitas, and emergence. Drawing on case studies from my other research, including the recurrent circulation of W. H. Auden’s “September 1, 1939” and poems by Maya Angelou following political violence, elections, and national trauma, I demonstrate how standard analytics can misread ritualized circulation as persuasion, consensus, or political momentum. In these cases, patterned spikes in circulation function less as evidence of emergent publics than as mechanisms of affective containment, ethical reassurance, and communal self-recognition. Crucially, ritual analytics foregrounds care as both an object of study and a methodological obligation. Analyzing ritualized responses to war, disaster, and political violence raises ethical concerns around extractive data practices, the quantification of grief, and the right to be forgotten. By incorporating best practices from cultural analytics, digital anthropology, and critical digital humanities, ritual analytics offers a framework for studying emergency response that is methodologically rigorous, politically attentive, and ethically responsible. By reframing circulation as ritual rather than discourse, this paper contributes to digital humanities conversations about care, response, and transformation, arguing that understanding how publics ritualize emergency is essential to imagining more genuinely emergent cultural and political futures.

Digital humanists are often asked to respond to emerging technology challenges in higher education and open publishing. Today, the focus is on AI. While many digital humanists offer a critical AI perspective, we will offer a framework for situating a legal perspective alongside humanities methodologies to respond to an urgent need to protect open knowledge. We discuss a policy agenda that recognizes education as a public good, pushing back against commercialization and commodification, against a rapidly changing political, economic, and technological background. Specifically, we focus on how to shift from a reactive posture to asserting a positive agenda of professional practices, institutional policy, and advocacy work. This panel will be performed in a dialogue style to answer a series of questions from the perspective of two legal scholars, specializing in copyright and fair use, in conversation with a digital humanities scholar and a humanities librarian. Using the Driving OER Sustainability for Student Success (DOERS) Collaborative "Policy Priorities for Generative AI and Open Education" (Cross and Jacob 2025) and guided by a fair use policy framework, our team will respond to questions like “Who owns our knowledge?” (Moumita 2025). Questions will include: (1) If open scholarship and publishing is a digital humanities value, how should digital humanists value open scholarship and publishing at this time in AI?, (2) How can digital humanists engage with AI in a non corporate way?, (3) At a time when engaging with AI feels like an all or nothing decision, what are use cases for responsibly engaging with AI for a digital humanist?, (4) What are simple ways I can think about the use of creative commons licenses to protect my digital scholarship?, (5) In a debate about the use of AI in a time when digital content is up for grabs between technology companies and publishers, and users and librarians are forced into subscription deals, how can we center support for public knowledge? Through this dialogue, we move beyond reactive anxiety to proactive engagement: not whether digital humanists should engage with AI, but how to do so while safeguarding the open knowledge commons we have built and the future we want to create.

Trained on stolen data, running in data centers that are rapidly depleting the Earth’s natural resources, and enriching companies through an accelerating cycle of speculation and spending, AI is rightly subject to critique. But as technologists and scholars of digital culture, the participants in this panel argue that the vision of AI instantiated by corporate actors is not the only possible outcome for these algorithmic systems. At the newly created CUNY AI Lab (https://ailab.gc.cuny.edu/), members of the CUNY Graduate Center have been exploring alternative visions of AI grounded in open-source technology, transparent practices, classroom workflows grounded in local knowledge, and classroom approaches that allow space for refusal and critique. In this panel, we will present some of our experimental tools, such as AmigAI, designed to support Spanish-language pedagogy for both heritage and non-native learning contexts, Multilingual Transcription Suite, which generates transcripts from audio and video in dozens of languages, and Document OCR, which creates searchable text from historical texts and archival documents. We will detail our evolving technical infrastructure, which leverages the open-source OpenWebUI platform to share zero-retention policies and open-weight models that can be used for AI experimentation in CUNY classrooms. Even as we present our model for AI experimentation, we will take up some of the recent critiques of so-called “open-source” AI models – can such systems rightly be called “open-source,” or do they instead represent the “open-washing” of problematic AI systems? How can and should DH practitioners balance their interest in playing with new technological tools with a responsibility to ground such play in skeptical critique? As we explore these questions and share our work, we hope to engage with the audience about both the dangers and possibilities of our present moment and to put the current exploration with AI in conversation with the long-standing tradition of critically informed, hands-on work of DH.

This panel brings together four digital humanities projects centered on Ukrainian language, culture, and history, each responding to a distinct but interrelated emergency: the erasure of Ukrainian material culture, the suppression of Ukrainian identity, and the ongoing challenge of making Ukrainian-language scholarship legible in a global research infrastructure. Taken together, the projects demonstrate how digital methods can serve as tools of cultural care, preservation, and resistance. The panel contributes to ACH 2026's focus on emergence/ia by asking how Ukrainian scholars and their collaborators are building new knowledge infrastructures from conditions of urgency. 1. Valeriia Korotkova: Vsesvit: From Data Collection to Data Storytelling Vsesvit (Всесвіт, The Universe) is Ukraine's oldest translation journal. Since its founding, it has played a crucial role in shaping and sustaining Ukrainian cultural identity throughout the Soviet period. This project documents the data collection and curation pipeline used to transform it into a digital database — its workflows, decisions, and challenges — and explores how structured metadata can tell stories, from journal geography and community building to shifts in editorial choices and canon formation. 2. Elizabeth Crim: Herbal Archives: Mapping Ecological Resistance in Ukraine This project examines Ukrainian medicinal herbal knowledge as a form of resistance to Russian ecocide through the digital indexing of Ukrainian herbal medicine using Geographic Information Systems (GIS). Additionally, this archive draws on Ukrainian-language corpora like ГРАК and qualitative coding in NVivo to trace how herbs represent larger narratives of survival, migration, and linguistic isolation. This project relies on a family archive of Ukrainian-language newspaper clippings assembled by my grandparents in Ukraine and the United States after they immigrated in the 1990s. Reframing the human–plant relationship through digital tools allows us to reconsider the role of ecological memory in countering ecocide. 3. Alyssa Virker: The Testimony of Ukrainian Poetry Videos on the Map This project maps the vastly popular form of Ukrainian poetry on TikTok in order to study the ways in which Ukrainians use poetry, art, and the act of documentation as a form of testimony during Russia’s War against Ukraine. The map reveals regional and national responses to war and trauma. In addition to the map itself, this project includes visualizations that show correlations between texts, visuals, and regions. This talk shares the process of collecting poetry videos, mapping them, and then analyzing data to show how Ukrainians create social media poetry videos to document Russian acts of war but also to encourage community, unity, and resistance amongst Ukrainians. 4. Georgii Korotkov: Segmenting the Archive Building on the PESP project at Princeton, which trained a YOLOv5 model (a widely-used object detection architecture) on Russian-language journals, this project fine-tunes that baseline on Vsesvit (the journal introduced by Valeriia) while benchmarking it against the more recent YOLOv8 architecture. Because Ukrainian typography shares visual conventions with Russian material but differs in script and decorative traditions, direct weights transfer doesn't generalize. The expanded taxonomy supports downstream metadata extraction, transforming pdf files into usable material for researchers.