CNI Spring 2026 Membership Meeting

Access and discovery are at the heart of what academic libraries do.  With the Americans with Disabilities Act Title II’s requirement that public entities’ web and mobile content needs to meet Web Content Accessibility Guidelines (WCAG) 2.1 Level AA standards, libraries and institutions are challenged by the volume of digital content that needs to be remediated and understanding expectations and requirements that impact existing operations in a resource-constrained time. Many libraries are working with partners in legal, disability offices, and IT to explore and find solutions that scale.

This panel will explore these challenges from multiple perspectives and hear about approaches and pathways being pursued at a few institutions. Attendees will be invited to share examples of efforts their organizations are undertaking and their questions to help everyone learn more about this complicated but important area.

Open at Scale: Exploring the Value and Impact of the BTAA's Open Publishing Agreement Model
Maurice York and Jeffrey Spies (Big Ten Academic Alliance)
 
From 2021 to 2024, the libraries of the Big Ten Academic Alliance (BTAA) paved the way for a dramatic increase in open access publishing by Big Ten authors—from 38% to more than 86% open in library-negotiated agreements. This was accomplished by innovating the Open Publishing Agreement (OPA) model, a strategy centered on the author experience: no fees, no caps, and no hassle. This values-based approach has delivered significant benefits for author rights, widespread author adoption, and measurable increases in research impact. This presentation will describe the current effort to explore the network of data from Big Ten publications—citations, authors, disciplines, and institutions—to further analyze the impact of these agreements: lowering barriers, expanding opportunities for authors, expediting knowledge sharing, and advancing research. Utilizing network visualizations, the BTAA is examining indicators of impact, such as shifts in citation patterns and global engagement that are directly evident from this initiative. The presenters will also discuss their strategic negotiation approach, including practical elements they are developing, such as the assessment rubric, license agreement terms, and evidence metrics.

https://btaa.org/library/open-scholarship/strategy

Advancing Open Monograph Opportunities at UC: New Pathways for the Future
Lidia Uziel (University of California, Santa Barbara) and Miranda Bennett (California Digital Library, University of California)

In January 2026, the University of California (UC) Libraries released Advancing Open Monograph Opportunities at UC, a report that articulates a values-driven framework for advancing open access monograph publishing. In this session, speakers will present the report and share how its recommendations are being translated into practice through coordinated pilots across the UC system.

While journal-based open access (OA) models have matured, monographs (particularly in the arts, humanities, and social sciences) pose distinct challenges shaped by disciplinary norms, funding structures, and the publishing economics of long-form scholarship. The report responds to this complexity by advancing a model-agnostic, portfolio-based approach that recognizes the diversity of publishing traditions and avoids reliance on any single funding mechanism.

Four strategic directions anchor the framework: targeted investment in book processing charge-based initiatives aligned with institutional research and teaching priorities; expanded support for Diamond OA and free-to-read models that remove both author- and reader-facing fees while advancing bibliodiversity and multilingual scholarship; strengthened partnerships with university presses as trusted stewards of peer-reviewed work; and sustained investment in open, community-owned infrastructure that ensures discoverability, metadata quality, preservation, and long-term sustainability. These strategies are guided by shared scholarly values, including equity, fiscal responsibility, transparency, and community stewardship.

The session will include highlights and lessons learned from UC's implementation of this framework through systemwide pilots that open UC-authored monographs at no cost to authors, combine frontlist and backlist approaches, and provide predictable, scalable support for university presses. Together, the framework and pilots demonstrate how a large research university system is aligning values with action, offering practical insights for institutions seeking sustainable futures for open monograph publishing.

https://escholarship.org/uc/item/9r22k58w
https://osc.universityofcalifornia.edu/2026/01/advancing-open-monograph-opportunities-at-uc-new-pathways-for-the-future/
https://osc.universityofcalifornia.edu/2026/01/open-access-for-uc-authored-monographs/

Academic libraries are under increasing pressure to deliver fast, high‑quality virtual reference services amid rising demand, staffing constraints, and expectations for always‑available support. Generative artificial intelligence (AI) is frequently marketed as a scalable solution, often without sufficient evidence about where automation might genuinely help or where it may introduce new burdens or risks of suboptimal service. This project briefing details a multi-institutional study of nearly 35,000 anonymized transcripts using a structured natural language processing (NLP) framework. It shares lessons learned from methodological evolution, testing a spectrum of approaches from off-the-shelf interfaces (Microsoft Copilot, Gemini) to a custom-engineered API pipeline. By pairing human expertise with an auditable codebook, the framework generates AI-reasoning trails for every assigned code, allowing researchers to see exactly where model logic aligns with or diverges from professional judgment. The results highlight promising opportunities for responsible automation but also document the extensive iterative cleaning and error-correction cycles necessary to achieve reproducible results. By foregrounding these insights and the often-invisible labor required, the session will provide library leaders with a realistic roadmap for responsible AI integration, moving beyond theoretical promises to offer evidence-based guidance on staffing, workflow design, and long-term strategy in scaling AI-assisted virtual reference.

In January 2026, the University of Virginia Library conducted the first of many proposed hands-on exercises to test four major large language models' (Anthropic's Claude, OpenAI's ChatGPT, Google Gemini, and Microsoft CoPilot) suitability for the transcription and translation of handwritten documents from the University of Virginia Library's special collections. Thirty-two staff members from multiple disciplines were given standard prompts and rubrics for evaluation, divided into groups, and given an opportunity to select an item to scan and provide to each LLM for transcription, and, if necessary, translation. Time was provided at the end for groups to share their findings and key insights. This initial exercise will be used to calibrate prompts and rubrics for similar events that will be held to continue evaluation of LLMs for processing of collections, and to inform faculty, students, and researchers on the effective use of these tools.

 https://library.virginia.edu/news/2026/gamechanger-can-ai-accurately-transcribe-primary-source-documents

Thousands of institutions have adopted the International Image Interoperability Framework (IIIF) to provide access to digitized collections, yet the content served through these APIs, including hundreds of thousands of maps, remains largely undiscoverable by location and difficult to use across institutional boundaries. Allmaps is an open source ecosystem that enables anyone to curate, georeference, and explore these resources without requiring GIS expertise or specialized infrastructure. In 2025, following the termination of a National Endowment for the Humanities grant, the IIIF Consortium formalized a tiered partnership with Allmaps to sustain its development. This is intended to be the first in a broader effort to help institutions realize the collective value of the content their shared infrastructure makes accessible via IIIF. This briefing will present the strategic rationale for this new consortial model from both the IIIF Consortium and Yale University Library, an early Allmaps Innovator. The session will explore how the partnership delivers tangible benefits to the community, from lowering barriers to georeferencing and cross-collection map discovery, to providing institutions with crowdsourcing tools, integration services, and governance participation without requiring specialized GIS infrastructure. Discussion will focus on how consortial organizations can move beyond standards-setting to actively steward and unlock the value of the content communities have collectively invested in digitizing.

https://allmaps.org/iiif-partnership

The University of Texas Libraries (UTL) entered into a charter program with ITHAKA in the fall of 2025 with the objective of testing Seeklight, JSTOR's artificial intelligence (AI)-based metadata generation tool. The goal for the first year of the charter was to test the quality of Seeklight's metadata output and its compatibility with multilingual and complex records. UTL evaluated the quality and changes over time over several months of testing, using born-digital and digitized records sourced from UTL's Benson Latin American Collection and Alexander Architectural Archive. The rubric used during this testing period is a four-point, subjective rating scale, which was developed by UTL's AI Metadata Creation Working Group during earlier tests with out-of-the box LLMs. This presentation will highlight our initial impressions of Seeklight's metadata output on a field-by-field basis. We will explore UTL's continued testing of Seeklight and its current ability to provide the subject knowledge and enhanced discoverability needs for archival records.

As artificial intelligence (AI) companies increasingly seek access to archival collections for model training, archival organizations face high-stakes decisions with limited precedent and no shared standard. The University of Virginia (UVA) Library has developed the UVA Archival AI Protocol (UVA AAIP), a practical framework grounded in a core rule: irreversible AI models do not get access to archival materials unless item-level provenance and meaningful attribution can be demonstrated, and the institution retains contractually enforceable control to stop further use. The Protocol distinguishes between retrieval-based AI systems—which keep source materials under institutional control and are generally permitted—and general-purpose model training, which absorbs knowledge into model weights irreversibly and is blocked by default. Built on three foundational pillars (provenance and attribution, donor and community responsibilities, and institutional control), the Protocol provides a decision framework for evaluating AI requests, sample contract clauses for deeds of gift and vendor agreements, minimum provenance standards for AI-generated citations, and a phased implementation plan designed for organizations of any size. This briefing addresses both the strategic rationale for the Protocol and the practical realities of putting it into action. Attendees will leave with a freely available adoption kit—including customizable templates, sample clauses, and implementation checklists—that any archival organization or memory institution can use to establish a principled, consistent position before the next AI partnership request arrives.

https://doi.org/10.18130/5dqf-9w86

The injustices embedded in the collecting and descriptive practices of libraries, museums, and archives are now widely recognized and have prompted many cultural institutions to pursue inclusive and reparative initiatives, such as harmful language statements and content warnings. Remediating outdated and offensive language in metadata is, however, a far more daunting task, especially at scale. While resources like the Inclusive Metadata Toolkit support informed reparative decision-making, identifying problematic terms remains a tedious and emotionally taxing process dependent on individual keyword searching. Consequently, many institutions either lack the time and resources to engage in this work or are constrained by the sheer volume of potentially harmful language present in metadata records. With the support of an internal library seed grant program, the J. Willard Marriott Library at the University of Utah  developed the Marriott Reparative Metadata Assessment Tool (MaRMAT), an open source, schema-agnostic, Python-based application for bulk metadata assessment. MaRMAT assesses tabular metadata against pre-curated and custom lexicons, generating a report flagging potentially harmful terminology by field, category, and context. By making large-scale assessment more accessible, MaRMAT empowers cultural heritage organizations to circumvent individual bias and advance equity in digital collections. MaRMAT's successful development also demonstrates the impact of small seed grant programs.

https://www.marmatproject.org/
https://github.com/marriott-library/MaRMAT

As generative artificial intelligence (AI) reshapes how students conduct research, academic libraries face a critical opportunity to redefine their value proposition. This session presents findings from two years of research with over 1,000 undergraduate and graduate students across disciplines, revealing that while AI chatbots have become essential cognitive partners for brainstorming and navigating complex research, students consistently turn to library resources for credibility verification and validation. The data demonstrates that students are thoughtful, skeptical consumers of AI who routinely cross-check AI-generated content against library sources, who view AI outputs, particularly citations and reasoning, with appropriate caution. Rather than displacing libraries, AI has elevated their role as the authoritative validator underlying AI-mediated workflows. This research offers evidence-based insights to inform strategic decisions around discovery systems, research support services, and resource access in an AI-integrated environment, positioning libraries not as competitors to AI tools but as essential stewards of trustworthy, credible knowledge in an era of information abundance.

There is always tension between an ideal and the reality on the ground. The ideal of a community and an organization working in partnership to ensure the sustainability of open source software was the impetus for the Lyrasis Organizational Home for Community Supported Technologies. Through the Organizational Home, Lyrasis serves as a fiscal agent and sponsor for open technologies for digital cultural heritage and scholarship. After more than ten years, the Organizational Home has evolved to include five very different community-supported programs: ArchivesSpace, CollectionSpace, DSpace, Fedora, and VIVO. In 2024, an analysis was conducted to assess the model’s effectiveness and identify changes needed to sustain its mission as a program partner, while ensuring the sustainability of these programs and itself. This effort was undertaken in collaboration with the staff and governance of the five programs and with the intention of full transparency into the process. The project briefing will present a review of the process, including its findings and outcomes with the Director of Community Supported Technologies at Lyrasis and Organizational Home program chair. The panel will explore whether Lyrasis met its goals (from the perspective of Lyrasis as well as from the programs) and what lessons might be shared and applied about the challenges and hopes for an open and sustainable technical infrastructure for scholarship and cultural heritage.

https://lyrasis.org/organizational-home/

Emerging international frameworks for disclosing artificial intelligence (AI) use are impacting how scholarship, libraries, universities, and publishers approach transparency. This session introduces the AI Transparency Declaration, the AI Disclosure (AID) Framework, and the broader landscape of Enacting AI disclosure in scholarly publishing through brief project presentations and an interactive conversation with the audience.

Sergio Santamarina (National University José Clemente Paz) will provide brief recorded comments and be available for questions via Zoom.


https://codeberg.org/ssantamarina/AI-Transparency-Declaration
https://crln.acrl.org/index.php/crlnews/article/view/26548/34482
https://openanthroresearch.org/index.php/oarr/preprint/view/435

Washington University (WashU) Libraries and the University of Kentucky Libraries have recently completed downloading millions of previously uploaded digitized files from the Internet Archive. This process was part of larger migration projects at both institutions: WashU's Newman Numismatic Portal's 76,000 assets are migrating to AM Quartex, and the University of Kentucky Libraries Kentucky Digital Newspaper Program's 86,000 assets were recently copied to a new local online discovery interface powered by Ex Libris Primo. This briefing will share background information on both institution's asset migration projects and details on the tools used, including the Internet Archive API, the command line interface from Internet Archive, and ChatGPT to assist with writing Python code to achieve this. Staff from both libraries will share how they independently developed workflows and discuss goals and desired outcomes that inspired the work, including preservation and access.

https://kdnp.uky.edu
https://journal.code4lib.org/articles/18510
https://nnp.wustl.edu/

In response to an increased and very real threat of cyberattack (as the United Kingdom recovers from the impact of the attack at the British Library), the Bodleian Libraries took the difficult decision to switch off 20 at-risk 'Specialist Digital Resources' in December 2024. The vocal comments from users demonstrated significant communities that were originally hidden. What looked on the surface like 'legacy resources' were still vital research tools, and it was important to make them available online again as quickly as possible. To do this, the Libraries had to be agile and creative but also strategic. To prevent similar cyber risks in the future, the Libraries developed infrastructure that was sustainable in the long term while also providing the familiar functionality and access to collections that researchers needed. This presentation will look at the lessons learned through the process—some predictable, some unexpected—now that 13 of the resources are available online again. It will also consider the role of library leaders to set strategic direction while balancing different and at times conflicting user, technological, and security needs.

The ARL/CNI AI Scenarios: AI-Influenced Futures was designed to guide readers through envisioning possible futures, ten years out, in which artificial intelligence (AI) has reshaped the landscape of research, higher education, and research libraries. In 2025, teams at Johns Hopkins University (JHU) libraries and the University of Delaware (UD) Library, Museums and Press deployed the scenarios to facilitate strategic planning conversations about AI in their institutions' libraries. These two institutions differ in significant ways, from their private/public status to their current centralized/distributed methods of supporting AI at the university. Yet, both planning processes leveraging the AI Scenarios surfaced similar strategic priorities, including the necessity for cross-institutional partnerships, staff upskilling, and cultural shifts within libraries around the knowledge, use, and implementation of AI. In both meetings, participants recognized various tensions, such as the value of openness versus embargoing of digitized collections, particularly as unique library materials hold unique value for large language model training.

This session will share how the JHU and UD teams ran internal strategic planning sessions using the AI Scenarios, detail the strategic themes that emerged during these discussions, and share how each is prioritizing and implementing findings at their institutions. During this session, the presenters will invite attendees to reflect on their own experiences in strategic planning, implementation, and navigation of organizational culture change in the context of AI. Attendees will have learned how the AI Scenarios were facilitated and analyzed at two very different institutions, providing the attendees with a strategy to either engage in the scenarios at their own institution or to learn from the unified findings from these two institutions.

Over the last thirty years, libraries, archives, and museums (LAMs) have built large-scale networks to make cultural heritage collections more widely searchable and available, including the Digital Public Library of America (DPLA) and its hubs, (e.g., The Historically Black Colleges and Universities Digital Collection and Mountain West Digital Library), as well as aggregators of cultural heritage collection descriptions (ArchiveGrid, Online Archive of California, Archives West), and other crowd-sourced efforts like Social Networks and Archival Context. All have engaged key challenges of standards, infrastructure, metadata, and of end users and their needs. However, most aggregations have struggled to define their value and to garner financial and organizational sustainability. Most are dependent on federal funders, foundations, and R1 institutions. Final reports from the National Finding Aid Network project, Allison-Bunnell's research, and the issues behind DPLA's transition all point to the critical gap in this area. Yet to date, there has been no cross-cutting analysis of these networks, the projects that confronted and overcame these challenges, and the successes and failures that will define the present and future of public knowledge and cultural heritage in the United States. The community can meet this existential moment with a thoughtful and informed reinvention of cultural heritage aggregation. This session will describe a new effort among key practitioners in this space and Ithaka S+R that proposes to find a new way forward for these large-scale networks. As the principals plan the project and seek out expertise on this economic and collective action challenge, they will use the session to engage with the CNI community on economic models, digital libraries, metadata, repositories, and special collections.

Higher education institutions and libraries are working to respond to ever-increasing cybersecurity threats and requirements. This panel will share experiences from a variety of libraries about responses to cybersecurity audits or incidents and their impact on operations. They will also discuss collaborations and proactive actions meant to help prepare for and mitigate future threats to our scholarly and archival data and systems.

This session will not be recorded.

From Strategy to Action: Building a Comprehensive GenAI Program in an Academic Library 
Jason Casden, Amanda Henley, and Rolando Rodriguez (University of North Carolina at Chapel Hill)

Over the past two years, the University of North Carolina (UNC) at Chapel Hill Library has developed and implemented a roadmap for supporting generative artificial intelligence (GenAI) needs on campus. These internal planning efforts positioned the Library to partner with campus-wide AI and IT leadership to design and deploy GenAI spaces, services, and tools. This project briefing will share the experience of establishing a GenAI portfolio across three themes: staff readiness, internal research, and public-facing services. Presenters will discuss the development of multiple individual projects, including launching the Library AI Studio, developing a GenAI instruction program, establishing a GenAI fellowship for undergraduates, incorporating GenAI into library workflows, and creating PromptLab—a sandbox environment where all UNC affiliates can safely explore and compare multiple large language models. The project-based strategy included both top-down directives from library leadership and bottom-up initiatives such as staff-led pilots. Attendees will leave with concrete examples of ways to respond quickly and strategically to campus needs emerging from this transformative technology.

https://library.unc.edu/ai/
https://library.unc.edu/ai/library-ai-studio/
https://library.unc.edu/ai/generative-ai-research/
https://library.unc.edu/ai/library-ai-studio/promptlab/

Building AI Readiness: A Multi-Pronged Approach at the University of Toronto Libraries 
Jacqueline Whyte Appleby (University of Toronto/Scholars Portal, Ontario Council of University Libraries) and Jess Whyte (University of Toronto)

Academic libraries stand at an inflection point with artificial intelligence (AI)—neither early enough to wait and see, nor late enough to rely on established best practices, which are still emerging and contested. The University of Toronto Libraries has developed a practical, multi-pronged strategy for building organizational AI readiness in this ambiguous middle ground in partnership with Scholars Portal, a shared infrastructure provider and digital preservation platform for Canadian academic libraries based at the University of Toronto. This presentation will discuss two complementary initiatives—one focused on collection readiness and one on Model Context Protocol development—that allowed the Library to explore infrastructure development, staff capacity building, and modes of user engagement in this space.

Collaborative Stewardship of Government Information Across Legacy and Born-Digital Formats 
Merrilee Proffitt (Democracy's Library US, Internet Archive)

This presentation will overview how the Internet Archive, Internet Archive Canada, and the Institute of Governmental Studies Library at the University of California (UC) Berkeley have collaborated to preserve and provide open access to government records and publications across legacy analog formats (books, paper records, microforms) and contemporary born-digital content. Drawing on efforts including the digitization of important federal materials in both Canada and the United States, the US End of Term web archive, and the LoCal Digitization Project's ongoing unification of dispersed local government documents, the session will explore how cross-institutional stewardship can enable new forms of research and public use. Beyond capture, the session focuses on designing collections and services responsive to researcher workflows, including longitudinal and cross-jurisdictional analysis, improved discovery, bulk access, and computational use. It will discuss prioritizing collections, metadata strategies, and infrastructure decisions. Because government information and publications are widely and redundantly held across CNI member institutions, the session will highlight opportunities for coordinated digitization, aggregation, and service development that extend the impact of work already underway. The presentation was developed in collaboration with Kathryn Stine (Digitization Project Planner, Institute of Governmental Studies Library, UC Berkeley), Kris Kasianovitz (Director, Institute of Governmental Studies Library, UC Berkeley), and Andrea Mills (Executive Director, Internet Archive Canada).

https://archive.org/details/democracys-library
https://igs.berkeley.edu/library/ca-local-documents-digitization-project
 
Building a Resilient Ecosystem for Publicly-Funded Research Data
Kristi Holmes (Northwestern University)

The Center for Open Science, with support from the Robert Wood Johnson Foundation, is leading a community-driven initiative to develop a strategic plan to ensure the long-term preservation, accessibility, and usability of publicly funded scientific data. This project emerged in response to the 2025 sudden removal of public data from multiple federal agency websites, highlighting critical vulnerabilities in scientific data infrastructure. The initiative convenes experts across research, policy, and data infrastructure to coordinate approaches that strengthen and sustain access to data generated through federal funding. The project aims to complement existing efforts by developing a framework for long-term stewardship of federally funded scientific data. Focus areas include monitoring at-risk repositories, ensuring the resilience of data repositories, promoting repository sustainability and resilience, and developing an outreach and advocacy framework to raise awareness among researchers, funders, policymakers, and the public. This effort aims to inform community-wide data stewardship practices and support the broader movement toward transparent and sustainable research data management.

https://www.cos.io/ensuring-preservation-accessibility-usability-of-public-data

Documenting Disruption: Collecting and Communicating About Rapid Change in the Scientific Enterprise (Lightning Talk)
Trevor Owens (American Institute of Physics)

Amid rapid policy and funding shifts under the second Trump Administration, the American Institute of Physics is leading efforts to document and preserve records and data of disruption across the US physical science enterprise. This talk highlights how an interdisciplinary team of historians, librarians, archivists, and social scientists is building infrastructures to capture contemporary change through story collecting, data analysis, and digital preservation. These initiatives aim not only to inform future scholarship but to empower communities today to navigate and shape scientific futures. We invite collaborators from across the CNI community to collaborate and help document, preserve, and explore change and resilience in the scientific enterprise.

https://www.aip.org/library/ex-libris-universum/physical-science-careers-disrupted
https://www.aip.org/statistics/impacts-of-restrictions-on-federal-grant-funding-in-physics-and-astronomy-graduate-programs
https://www.aip.org/library/ex-libris-universum/adding-your-photos-to-the-story-of-science-is-now-easier-than-ever
https://www.aip.org/library/ex-libris-universum/join-us-in-documenting-celebrating-womens-contributions-to-the-physical-sciences

Institutional Repositories (IRs) were envisioned in academic libraries as comprehensive platforms capable of supporting open access publishing, research outputs, and digitized special collections within a single system. In practice, however, most IRs and related digital asset management software (DAMS) were not designed around archival principles or data structures, often resulting in duplicative metadata creation, digital materials separated from their archival context, and workflows that could not fully leverage archival efficiencies at scale. Over time, these design limitations have carried significant organizational consequences, contributing not only to miscommunication between digital and archival processes within libraries but also to the development of multiple publicly funded repository platforms with overlapping scopes in New York State and beyond. This presentation examines these systemic challenges and argues that the emergence of IIIF, combined with the IMLS-funded ArcLight Integration Project and the Delivering Archives and Digital Objects Conceptual Model (DadoCM), offers a path forward. By clearly articulating archival requirements and enabling greater interoperability, academic libraries can move toward repository architectures that reduce redundancy, better leverage automation, and more fully realize the broad institutional vision that originally motivated IR development.

PLEASE NOTE: This session is 30 minutes (1-1:30 p.m.)

https://archives.albany.edu/arclight_integration/
https://dadocm.github.io/

Scoring AI for Accessibility: A Rubric-Based Framework
Sarah Cogley and Stacy Snyder (University at Buffalo, SUNY)

This briefing explores a comprehensive evaluation of generative AI tools for creating alternative text and long descriptions for digital collections at the University at Buffalo Libraries. Prompted by the need to address extensive accessibility remediation as required by the Americans with Disabilities Act Title II regulations, the project team analyzed outputs from three AI tools using a rubric designed to assess three criteria: factual accuracy and correctness, relevance and task completion, and clarity and communication quality. Findings showed limitations of current AI technologies, such as hallucination, omission of key visual elements, and cultural insensitivity. The briefing will discuss challenges, such as how to evaluate tools that are emerging and dynamic, establishing guidelines and best practices for accessible metadata in digital collections that contain diverse content and format types, and how to integrate AI in digital collections workflows in a scalable and sustainable way. The presenters will highlight lessons applicable across institutions, including the importance of cross-unit collaboration with colleagues in user experience and accessibility, and will share the rubric, workflows, and project documentation.

Using Artificial Intelligence to Extract and Understand Cultural Heritage Materials
Paul Gallagher (Western Michigan University)

Much of our cultural legacy is hidden. Despite decades of effort to convert paper documents to electronic form, many issues still impact users' ability to discover content. Traditional optical character recognition only works so well; handwritten documents need manual transcription; and digital content platforms don't always provide researchers with meaningful ways to interact with historic content. In this presentation, learn how one library is working with emerging artificial intelligence (AI) models to extract text from heavily degraded documents and historic handwriting, using modern "vibe-based" application development to present cultural legacy items in a new way. Learn what tools are available, how they are best used, and how declining technical barriers will make this more accessible to information professionals. Gain a deeper understanding of powerful uses of AI beyond chat models, and how these approaches may benefit your own organization. Drawing on lessons from a recent project, this session is designed for librarians, archivists, and digital humanities practitioners interested in practical applications of AI for cultural heritage materials—regardless of technical background.

Expanding Access to Historic Scanned Documents Using R's Tesseract Package
Adelynn Shirts and David Advent (Utah State University)

Utah State University's Institutional Repository, DigitalCommons@USU, hosts over 100,000 PDF documents, many of which were originally printed pre-1975 and then later scanned. As such, they lack embedded text layers, rendering them inaccessible to screen readers without additional processing. A scalable pipeline was built to identify documents lacking embedded text and perform optical character recognition (OCR), making the content accessible to screen readers. Two preprocessing functions deskew, denoise, and enhance document clarity prior to performing OCR. Dictionary coverage from light and heavy preprocessing functions were compared: light preprocessing was computationally faster but resulted in less dictionary coverage, while heavy preprocessing added a modest amount of time and increased dictionary coverage slightly. After evaluating outputs, it was determined that the dictionary coverage of documents lacking embedded text layers were similar to those containing embedded text layers. While this doesn't make documents exactly compliant with Americans with Disabilities Act standards, it is an important first step in working towards accessibility for older publications, especially considering the open source nature of the code and process.
https://github.com/ashirts/Expanding-Access-to-Historic-Scanned-Documents

Improving Accessibility and Discoverability Utilizing Open Source Models in a Novel Modular Design
Brian McBride, Harish Maringanti, and Bohan Zhu (University of Utah)

University libraries are under growing pressure to expand access and improve discovery while meeting new accessibility expectations for  digital collections. At the University of Utah's J. Willard Marriott Library, the digital infrastructure development team is building a flexible, modular workflow to help bring large audio-visual (AV) collections into alignment with the Department of Justice accessibility requirements at scale. The platform orchestrates open source speech-to-text and language models to generate time-aligned transcripts and captions, structured segmentation, word clouds, entity recognition, and descriptive metadata that improves both compliance and discoverability. The session will highlight the  implementation approach, early results, and lessons learned including human review checkpoints, staff support and buy-in, provenance and auditability, and how adaptable workflows are being designed  as models and standards evolve. The session will also focus on practical strategies other institutions can reuse to accelerate accessible AV delivery without locking into a single vendor or toolchain and the team’s future development plans for supporting other formats, including images, PDFs, and other formats.

Strategies for Responsible AI in Manuscript Transcription (Lightning Talk)
Sara Brumfield (FromThePage)

FromThePage is a crowdsourcing platform for archives and libraries where volunteers transcribe, index, and describe historic documents. This talk will overview how the platform and community are making decisions that make the use of artificial intelligence (AI) in historical document transcription transparent, optional, and tentative, including topics such as:
- Optional usage of AI by transcribers and institutions
- Surfacing and logging use of AI drafts
- Provenance in exports showing both AI and human contributions
- Detecting unauthorized use of AI
- Measuring accuracy

http://www.fromthepage.com