Wednesday, May 6, 9AM CST/10AM EST to 3PM CST/4PM EST

The inaugural SEC Libraries AI Summit brings together an extraordinary range of voices from across academic libraries to explore how artificial intelligence is actively reshaping our work - from instruction and research support to metadata, digital scholarship, and copyright. These sessions move beyond theory and hype to showcase practical, real-world applications: teaching students to critically engage with AI-generated content, integrating AI into research consultations and data workflows, experimenting with metadata creation at scale, and navigating the rapidly evolving legal and ethical landscape. Whether you are just beginning to explore AI or already implementing tools and services, this summit offers concrete strategies, cautionary insights, and innovative models you can adapt to your own institution. To deepen the conversation, each thematic block (Copyright, Policy & Strategy, Digital Scholarship, Data & Emerging Technologies, Teaching & AI Literacy, AI Tools in Action, and Cataloging, Metadata & Archival Solutions) will conclude with a live Q&A, giving attendees the opportunity to engage directly with all speakers in that section.

Designed with flexibility in mind, the summit is a free, come-and-go event running from 9:00AM to 3PM Central 10:00 AM to 4:00 PM Eastern on Wednesday, May 6, making it easy to join for the sessions most relevant to your interests and schedule. Drop in for a single lightning talk or stay for the full experience - either way, you'll leave with fresh ideas, new connections, and a clearer sense of how libraries can lead in this rapidly evolving space.

We are committed to fostering a respectful and welcoming environment for all participants. We ask that all attendees engaged in discussions with professionalism, civility, and an openness to diverse perspectives.

9AM CST/10AM EST
Welcoming Remarks

  • David Banush, Dean of University Libraries at the University of South Carolina

9:10AM CST/10:10AM to 10:10AM CST/11:10AM EST
Copyright, Policy & Strategy

  • Helping More Users and/or Training My Replacement? Impressions from Testing a Scholarly Communication Chatbot
    Dylan Martin, Social Sciences & Copyright Librarian, University of Missouri Libraries
    To support a strategic goal of sustainable revitalization while expanding user-facing services, this project explored whether a chatbot could effectively answer common copyright and scholarly communication questions. The ScholarlyPub AI Chatbot was developed using Google AI Studio and Gemini Academic V2.0 and tested from an end-user perspective. This session examines how well the chatbot performed across typical inquiries and reflects on broader implications, including questions of liability and professional displacement. It offers a candid assessment of the opportunities and risks associated with implementing specialized AI tools in library contexts.
  • Designing AI Exploration for Maximum Learning: A Practical Framework for Academic Libraries
    Dhanushka Samarakoon, Director of Technology Strategy & Planning, University of Miami Libraries
    The rapid emergence of artificial intelligence presents academic libraries with both opportunity and uncertainty. This session introduces a learning-first approach to AI exploration developed at the University of Miami Libraries, designed to build internal capacity and shared understanding before committing to production-level services. Beginning with a proof-of-concept AI-powered search interface, the project provided hands-on experience with generative AI APIs, prompt engineering, and knowledge retrieval systems. Key insights included the volatility of the AI ecosystem, model inconsistency, API instability, and long-term maintenance challenges. These lessons informed a strategic pivot toward managed AI platforms that prioritize accessibility, sustainability, and institutional comprehension. The session presents a practical framework for designing AI initiatives that maximize learning, identify pivot points, and build organizational buy-in.
  • Subject Librarianship and AI Competencies: Librarian Perspectives
    Valrie Minson, Associate Dean for Research, University of Florida Libraries; Laura Spears, Director of Assessment, University of Florida Libraries
    This session addresses the growing need to equip both subject and functional librarians with essential AI competencies. Drawing on a 2025 survey of academic library faculty, the presenters explore how librarians perceive the distribution of AI-related skills across roles. Competencies discussed include understanding foundational technologies, critically evaluating digital information, ensuring ethical accountability, and optimizing AI-generated content. The session offers practical strategies for building and assessing these competencies and fostering a collaborative environment that supports ongoing professional development in AI.
  • Copyright & AI: Who Owns Creative Works in the Age of Machines?
    Sierra Whitfield, Copyright/Fair Use Librarian, Texas A&M Libraries
    As artificial intelligence becomes increasingly integrated into research and creative processes, questions of authorship and ownership have taken on new urgency. This session provides a practical overview of how U.S. copyright law applies to AI-generated and AI-assisted works. Current legal interpretations reaffirm that copyright protection is grounded in human creativity, positioning AI as a tool rather than an author. The presentation explores when AI-assisted works may qualify for protection, the risks associated with relying solely on AI-generated content, and strategies for preserving human authorship within evolving scholarly practices.
  • Before You Sign: Teaching Copyright in the Age of AI
    Amie Freeman, Assistant Head of Acquisitions & Scholarly Communication, University of South Carolina Libraries
    Copyright transfer is a routine yet often misunderstood aspect of scholarly publishing. This session introduces a workshop designed to help researchers better understand publishing agreements and manage their rights, particularly in the context of artificial intelligence. As AI systems increasingly draw on scholarly outputs for training, questions surrounding authorship, reuse, and ownership have become more complex. The workshop addresses these issues through practical examples, common misconceptions, and interactive activities. Attendees will gain ideas for developing similar instructional programs that empower researchers to make informed decisions about their work.

10:10AM CST/11:10AM EST to 11:10AM CST/12:10PM EST
Digital Scholarship, Data & Emerging Technologies

  • Multidisciplinary Research Proposal Enhancements through Large Language Models
    James Creel, Director of Scholarly Communications and Data Services, Texas A&M University Libraries; Ethel Meija, Senior Data Analyst, Texas A&M University Libraries; Daniel Xiao, Digital Resources Librarian, Texas A&M University Libraries
    Research Information Management Systems (RIMS) house rich, structured data on faculty expertise, yet they remain underutilized for fostering multidisciplinary collaborations in grants, mentoring, and related initiatives. The Office of Scholarly Communications at Texas A&M is piloting an approach that leverages profiles from the Scholars@TAMU RIMS, combined with data from grant calls and institutional funding programs, as inputs to large language models to surface unexpected collaboration opportunities. In partnership with researchers, the team is evaluating and refining this methodology in practice. This presentation shares preliminary findings, along with code and supporting resources to facilitate replication and further study.
  • Can a Custom GPT Improve First-Pass Data Management Plan Review?
    Stacy Winchester, Research Data Librarian, University of South Carolina Libraries
    Many data librarians offer Data Management Plan (DMP) review services to help principal investigators ensure compliance with funder requirements and alignment with FAIR principles. This session explores the development and testing of a custom GPT-based tool designed to assist with initial DMP review. By flagging common omissions and potential issues, the tool aims to streamline the review process and support data librarians in delivering efficient and effective services.
  • Managing Massive PDF Remediation Projects with AI Support
    Amie Freeman, Assistant Head of Acquisitions & Scholarly Communication, University of South Carolina Libraries; Lance DuPre, Digital Technologies Development Librarian, University of South Carolina Libraries
    The responsibility to provide accessible digital content remains ongoing, and libraries face significant challenges in remediating large volumes of legacy PDF files. This session explores the use of an AI-powered tool to support PDF remediation and accessibility compliance. Faced with the need to remediate tens of thousands of documents, the team evaluated multiple AI-based accessibility solutions before selecting a PDF remediation system developed through the Arizona State University AI Cloud Innovation Center using AWS. The presentation outlines implementation logistics, practical considerations, and lessons learned. Attendees will gain insight into how AI can reduce staff burden, improve efficiency, and support accessibility compliance at scale.
  • Teaching AI-Enhanced Podcasting Across Disciplines
    Dr. Chelsy Hooper, Instructional Technologies Coordinator, Auburn University Libraries
    Podcasting offers a dynamic platform for students to share research findings, conduct expert interviews, and synthesize complex ideas. This session explores an interdisciplinary approach to teaching podcasting within a library setting, supporting students across fields such as engineering, journalism, history, and animal science. By integrating AI-enhanced audio tools, students can achieve studio-quality sound while meeting editing and accessibility needs. Tools such as Adobe Podcast, Adobe Audition, and Adobe Express support audio production and visual design. The session highlights instructional strategies, student engagement, and the development of transferable skills that extend beyond the classroom into professional and creative contexts.
  • Dr. ClaudeCode, or How I Learned to Stop Worrying and Love to VibeCode
    Heather Heckman, Associate Dean for Information Resources & Technologies, University of South Carolina Libraries
    I will discuss how I used ClaudeCode to complete in a matter of hours an empirical humanities paper I have been working on for years. I will argue that this has profound implications for how academic libraries structure and prioritize our services. In digital services and data services, librarians have helped to mitigate technical knowledge deficits in the research community. Arguably, subject area expertise is now the limiting factor for empirical and computational research.

11:10AM CST/12:10PM EST to 12PM CST/1PM EST
Lunch Break

12PM CST/1PM EST to 1PM CST/2PM EST
Teaching & AI Literacy

  • Real or Robot: An AI Literacy Microprogram
    Peyton Consani, Information Literacy Librarian, Southeast Missouri State University; Olivia Kays, Academic Library Fellow, University of Missouri-Kansas City Libraries; Dani Wellemeyer, Head of Outreach & Engagement, University of Missouri-Kansas City Libraries
    Seeking ways to connect with our students on the topic of AI literacy outside of the library instruction classroom, we developed our "Real or Robot" microprogram. It opens conversations that guide students to view the library as a place where experts teach about information creation and effective and ethical uses of generative AI. Interactions reveal students' ability to identify some types of AI-generated content; more difficult questions keep them engaged in learning. Even this small-scale use of AI in teaching assesses understanding, informs development of larger projects, and aligns with our library's work to establish its role in the campus landscape for AI education.
  • Fast, Easy...and False? Not Falling for AI's Hallucinations
    Ripley Owens, Library Specialist II, Texas A&M Medical Sciences Library; Amy Martin-Klumpp, Research & Education Librarian, Texas A&M Medical Sciences Library
    AI tools are appealing because they feel fast, embedded, and endlessly accessible, but that convenience comes with a growing problem: hallucinated information that sounds correct but falls apart under scrutiny. At our institution, AI-generated references frequently arrive through our Get It For Me service, creating extra work as staff cancel requests and untangle citations that never existed. We're also increasingly seeing students and faculty rely on AI-generated content without recognizing how hallucinated data or references can quietly distort their findings. In this lightning talk, we'll share how we've learned to recognize AI-generated citations and information, why AI often presents misinformation with confidence, and how we guide users to reframe AI as a starting point in the information-gathering process rather than an end product. A key insight is that while AI may be useful, its generated content cannot be assumed to be fully accurate and does not eliminate the need for critical thinking.
  • Lean on Me: Supporting Critical Thinking Using Databases and AI
    Lauren A. Camarillo, Research & Education Librarian, Texas A&M Medical Sciences Library; Amy Martin-Klumpp, Research & Education Librarian, Texas A&M Medical Sciences Library
    "Leaning in" to AI means maintaining professional agency in navigating when it's used (if at all), learning how to use it effectively, and maximizing its potential. As vendors increasingly integrate AI capabilities within library databases, librarians are engaging with these tools whether they intend to or not. Our team of medical librarians leans into organic opportunities that highlight AI as both a practical tool (not a solution) and a cautionary liability. Strategically using teachable moments enables librarians to emphasize critical thinking within the inquiry process. Key takeaways include identifying teachable moments and understanding algorithms in AI and databases.
  • Expanding Scholar Services Through AI Research Appointments
    Kristina Bloch, Engineering Librarian, University of Louisville
    Academic libraries are navigating growing demand for AI support in research while balancing ethical use, disciplinary norms, and researcher uncertainty. This lightning talk introduces the expansion of a Scholar Services program to include AI Scholar Research Appointments, a structured consultation model that integrates generative AI into the research lifecycle rather than treating it as a standalone tool. These appointments focus on search strategy development, transparency, ethical decision-making, and critical evaluation of AI outputs. Early findings suggest increased researcher confidence, clearer boundaries for appropriate AI use, and deeper librarian–faculty partnerships. Attendees will gain a replicable model for integrating AI into research services while maintaining scholarly rigor and alignment with information literacy frameworks.
  • From Buzzwords to Buy-In: Building AI Workshops That Work in Academic Libraries
    Dr. Deborah Lee, Director of Research Impact & AI Strategy, Mississippi State University Libraries
    As interest in generative AI accelerates across higher education, academic libraries face both an opportunity and a challenge: moving programming beyond hype toward informed, ethical, and practical use. This lightning talk presents a collaborative model for building AI-themed workshops and instruction through partnerships with key campus stakeholders. Drawing on real-world workshop development, the session highlights how AI programming can align with institutional priorities, promote digital literacy, and extend the library's impact. Attendees will gain practical insights into designing AI initiatives that leverage campus relationships to create sustainable, mission-driven learning opportunities.

1PM CST/2PM EST to 2PM CST/3PM EST
AI Tools in Action

  • AI as Research Companion: Enhancing Undergraduate Engagement with Archival Materials
    Sheila Devaney, Humanities & Social Sciences Librarian, University of Georgia Libraries Megan Palmer, STEM Librarian, University of Georgia Libraries
    NotebookLM is a unique AI-powered research tool that relies on materials the user uploads to explain, summarize, compare, and generate content. UGA librarians saw this as an opportunity to explore how AI might support undergraduate engagement with archival and other primary sources. They compiled a curated set of materials related to the Brooklyn Cemetery in Athens, Georgia - including cemetery records, newspaper coverage, and oral histories - which were then uploaded into the system. The chat feature allows students to identify connections and recurring themes across sources, trace shifts in language over time, and compare official narratives with community perspectives. Rather than replacing close reading, NotebookLM functions as a scaffold for discovery, helping students surface research questions and contextual patterns while remaining grounded in the archival record itself. At the same time, the tool emphasizes critical engagement: students are encouraged to interrogate the biases embedded in both the archival materials and the AI-generated synthesis, and to situate their findings within broader historical scholarship.
  • Using NotebookLM and Google Gemini Pro for Policy Analysis
    Elizabeth White, Humanities & Social Sciences Librarian, Univerity of Georgia Libraries
    This session explores how AI tools such as NotebookLM and Google Gemini Pro can be used to discover public datasets and assist students in analyzing policy documents. These tools support more efficient data discovery and provide new avenues for engaging with policy-related research materials.
  • Initial Insights on Teaching with Local Generative AI Language Models
    Dr. Rachel L. DuBose, Data Science Librarian, University of Alabama Libraries; Dr. Vincent Scalfani, Director of Research Computing Services, University of Alabama Libraries
    Numerous academic libraries offer instruction and guidance on the appropriate use of service-based generative AI tools such as ChatGPT in education and scholarship. In addition to teaching users about these tools, libraries also have an opportunity to teach users how to run generative AI language models locally - that is, setting up language model inference on local hardware. Running models locally can offer unique advantages such as offline functionality, increased privacy, and customization. This session provides an overview of popular options for running local models, along with initial lessons learned from developing tutorials and instructional approaches.
  • Reproducibility and Utility of AI Search Tools in Evidence Synthesis: A Comparative Study of Consensus and Perplexity
    Dani LaPreze, Evidence Review & Synthesis Librarian, Texas A&M Libraries
    The rapidly evolving landscape of AI search engines offers the potential to streamline evidence synthesis projects by generating searches and providing immediate, synthesized answers within a single platform. However, the integrity of evidence-based practice depends on two critical elements: reproducibility of results and relevance of retrieved sources. This study evaluates two AI platforms by assessing citation accuracy and result replicability. Preliminary findings highlight limitations and risks in AI-assisted searching for evidence synthesis projects. The session provides practical guidance for librarians seeking to responsibly integrate AI tools into workflows while maintaining methodological rigor and transparency.
  • Using NotebookLM as an Interactive Repository for Evidence Synthesis Standards
    Esmeralda Garcia Rogers, Evidence Review & Synthesis Librarian, Texas A&M Libraries
    Navigating the diverse standards for evidence synthesis - from JBI to specialized integrative review frameworks - can create a significant cognitive burden for librarians. Google NotebookLM can be used to transform static documents into interactive methodological repositories. By uploading standards such as the JBI Manual, PDFs become part of a conversational interface that supports exploration and application. This workflow also enables the rapid generation of customized researcher checklists, improving efficiency and accuracy in consultations. Attendees will learn how to leverage AI as an "external memory" for evidence synthesis guidelines while maintaining transparency through appropriate AI disclaimers in researcher deliverables.

2PM CST/3PM EST to 3PM CST/4PM EST
Cataloging, Metadata, & Archival Solutions

  • AI in Action: Practical Experiments in Cataloging at the University of Miami Libraries
    Jose Vila, Unit Lead for Interdisciplinary Collections, University of Miami Libraries; Kristina Yost, Library Technician, University of Miami Libraries
    Academic libraries face increasing pressure to process growing and multilingual collections while maintaining high metadata quality standards. Cataloging units must manage persistent backlogs, large-scale data cleanup, and repetitive normalization tasks, all while meeting rising productivity expectations with limited staffing. This creates a constant tension between scale and quality in cataloging workflows. At the University of Miami Libraries, artificial intelligence is being explored as a tool to support, rather than replace, catalogers' expertise in addressing these challenges. This presentation highlights practical case studies where AI has been applied to metadata creation, data remediation, and transliteration for record creation. Examples include generating AI-based summaries for over 1,000 marine science theses to improve discovery, batch normalization of item descriptions, comparison of generative AI tools for bibliographic record creation, and experiments in Arabic transliteration. The session shares what worked, what failed, and where human review remains essential, offering attendees concrete ideas for low-risk AI experimentation that improves efficiency while preserving professional standards.
  • From Pilot to Production: Scaling AI-Assisted Metadata Creation for Legacy Musical Collections at University of Texas at Austin Libraries
    Kayode (Kay) Ishola, Automation & Integration Librarian, University of Texas at Austin Libraries
    The University of Texas at Austin Libraries is using artificial intelligence to support metadata creation for large legacy music collections, including more than 300,000 CDs and LPs. This presentation traces the evolution of an AI-assisted cataloging project from early experimentation to a validated workflow that incorporates human review. It explores how AI is used to generate descriptive metadata and match records to external databases, what has been learned through testing, and how the project is being prepared for operational scale. Future directions include exploring robotics to automate scanning as part of an end-to-end workflow. Attendees will gain practical insights into scaling AI-driven metadata creation while maintaining quality and consistency.
  • Digital Collections at USC Explores JSTOR Seeklight AI for Metadata and Transcription
    Katie Hoskins, Digital Collections Librarian, University of South Carolina Libraries
    Metadata work is time-intensive, and transcription even more so, yet discovery and accessibility of digitized archival materials depend on accurate description and full text. Since October, the University of South Carolina's Digital Collections team has participated in JSTOR's Digital Stewardship charter program, testing and refining Seeklight, an AI tool for generating metadata and transcripts. The team has integrated Seeklight into descriptive workflows for photographs and manuscripts, and the tool has become central to the department's strategy for achieving accessibility of handwritten historical documents. This presentation shares findings on the tool's accuracy, efficiency, and impact on metadata and transcription practices.
  • Let's Stop Talking and Start Building: A Call for Collective Action with AI in Academic Libraries
    Glenn Bunton, Data Visualization/Gen AI Librarian, University of South Carolina Libraries
    , Agentic AI is here, and academic libraries risk falling behind - not because of a lack of capacity, but because efforts remain fragmented and overly deliberative. This session challenges libraries to move from isolated experimentation to coordinated action. It proposes a focused, collaborative workshop model organized around core functional areas such as cataloging/metadata, research, archives, acquisitions, and digital services. Each group would identify shared AI opportunities and propose scalable solutions, including generative and agentic AI applications. The key takeaway is a roadmap for shifting from parallel exploration to collective construction, advancing a more unified and productive future for academic libraries.
  • Assessing Large Language Models: Architectural Archive Metadata and Transcription
    Hannah Moutran, AI & MP Library Specialist, University of Texas at Austin Libraries; Devon Murphy, Metadata Analyst, University of Texas at Austin Libraries; Karina Sánchez, Digital Scholarship Librarian, University of Texas at Austin Libraries; Katie Pierce Meyer, Head of Architectural Collections, University of Texas at Austin Libraries; Josh Conrad, Digital Initiatives Archival Fellow, University of Texas at Austin Libraries
    This session examines the use of large language models in generating and enhancing metadata and transcription for architectural archival collections. The presenters explore how AI can support description and access for complex archival materials, with attention to evaluation, accuracy, and practical implementation in workflows.

Follow