Investing in AI: a review of recently funded IMLS projects
Libraries are steadily embracing artificial intelligence to transform how they manage, access, and disseminate information. I reviewed all 89 of the recently awarded IMLS grants (library services office) to see what people are building. Lots of fantastic projects! Congrats to everyone. But let’s look at what’s underway with AI. Sometimes there are questions of “how will libraries use AI?” Well, let’s see what’s being funded.
Note: also see my companion piece: Investing in Communities: a quick review of recently funded IMLS projects
Enhancing Information Discovery
AI is being used to improve information discovery. Northwestern University Libraries are developing an open-source semantic discovery tool that allows users to interact with collections through chat. Similarly, the University of Maryland iSchool is creating a user-centered chatbot to enhance exploration within library and archival collections. These AI tools aim to simplify and speed up the process of finding relevant information.
Automating Metadata and Cataloging
Libraries are also turning to AI to streamline metadata creation and cataloging. The University of Michigan and the University of North Texas are employing generative AI and large language models (LLMs) to automate the creation and evaluation of metadata. This automation intended to save time and increase accuracy and consistency.
Boosting Digital and Media Literacy
Educational initiatives are benefiting from AI integration as well. Carnegie Mellon University Libraries are enhancing algorithm, data, and media literacy through its Project on Open and Evolving Metaliterature. This project aims to equip learners with essential skills to navigate and understand the emerging and evolving digital landscape.
Creating Open Educational Resources
The University of Virginia Library is using generative AI to support the creation and adaptation of open educational resources (OER). By developing a community web portal, the project provides a collaborative space for librarians, educators, and AI practitioners to share knowledge and tools, democratizing access to high-quality educational materials. Additionally, Carnegie Mellon University (as noted above) is developing OER content too.
Building Collaborative Platforms
AI is also being used to build collaborative platforms that enhance community engagement and knowledge sharing. The University of Maryland’s project includes community-driven AI design, fostering a collaborative approach to developing AI tools that serve diverse user needs.
ETDs + LLMs
AI is proving to be a valuable asset in research assistance too. Virginia Tech’s University Libraries are integrating electronic theses and dissertations (ETDs) with large language models (LLMs) to provide advanced information discovery and research trend analysis. This integration will help researchers find relevant studies and data more efficiently.
Fostering AI Literacy and Adoption within Libraries
Virginia Tech and the University of California Riverside’s Generative AI Incubator Program aims to create AI applications specific to library use, such as GenAI literacy and collection preservation. These effort will provide hands-on experience and training for library professionals.
The Future of Libraries and AI
As these projects demonstrate, libraries are actively integrating AI into their operations, from enhancing collection discoverability to expanding engagement. This grouping showcases a range of initiatives: some libraries are focusing on tool development and digital infrastructure, while others emphasize literacies, training, OERs, adoption efforts, and the more social aspects of AI. It's inspiring to see IMLS funding these innovative projects, leading the way for the future of libraries. These efforts not only improve accessibility and efficiency but also ensure that libraries remain vital, dynamic centers of knowledge and community interaction across today’s information landscape.
Quick Overview
Here is the quick outline of the eight projects and the deliverables:
1. University of Maryland iSchool
Abstract:
The University of Maryland iSchool will develop a user-centered chatbot using generative AI to enhance exploration and discovery within library and archival collections. The project includes community-driven AI design, making archival collections AI-ready, and comparing AI and non-AI query results.
Deliverables:
Development of a user-centered chatbot.
Community-driven AI design and evaluation.
Study on making archival collections AI-ready.
Comparative analysis of AI and non-AI query results.
Focus groups for testing and evaluation.
Generalizable approaches for AI integration in collections.
2. Northwestern University Libraries
Abstract:
Northwestern University Libraries will use generative AI to build an open-source semantic discovery tool for chatting with collections and an automated metadata augmentation tool.
Deliverables:
Development of an installable open-source semantic discovery product.
Implementation of an automated metadata tool.
Toolkits with research methods, examples, documentation, and software packages.
Integration and validation of tools in Northwestern’s Digital Collections.
Dissemination of findings through documentation and software packages.
3. Carnegie Mellon University Libraries
Abstract:
Carnegie Mellon University will enhance algorithm, data, and media literacy through the Project on Open and Evolving Metaliterature, targeting high school and above learners.
Deliverables:
Creation of open educational resources on algorithm, data, and media literacy.
Development of practical, peer-reviewed educational materials.
Integration with existing literacy frameworks and initiatives.
Dissemination of materials to educators and learners.
4. University of Michigan School of Information and ICPSR
Abstract:
The University of Michigan will explore using generative AI to aid data curation by drafting and evaluating metadata and managing sensitive data elements.
Deliverables:
Development and testing of GenAI tools for metadata curation.
Evaluation of GenAI tools’ effectiveness.
Integration of computational tools for managing sensitive data.
Feedback and user studies for tool improvement.
Documentation and dissemination of findings.
5. Virginia Tech’s University Libraries and UC Riverside
Abstract:
Virginia Tech and UC Riverside will launch a Generative AI Incubator Program to create AI applications for library use, focusing on literacy, collection preservation, and research.
Deliverables:
Development of GenAI literacy, collection preservation, and research applications.
Multi-day workshop and mentoring sessions.
Capstone projects demonstrating AI applications.
Training materials and workshops.
Dissemination of deliverables and project outcomes.
6. University of North Texas Department of Information Science and University Libraries
Abstract:
The University of North Texas will investigate using locally run large language models (LLMs) to assist in cataloging digital and print library resources.
Deliverables:
Development of LLM-based models for accurate cataloging.
Integration of LLMs into cataloging procedures.
Methodologies and code for LLM-based models.
Analysis and reports on LLM applications.
Dissemination of findings and best practices.
7. University of Virginia Library
Abstract:
The University of Virginia Library will create a community web portal using generative AI to support the creation and adaptation of open educational resources (OER).
Deliverables:
Development of a community web portal.
Integration of generative AI technologies.
Creation and adaptation of OER.
Resources including repository, discussion forum, and events calendar.
Collaboration tools for librarians, educators, AI practitioners, and students.
8. Virginia Tech University Libraries and Department of Computer Science
Abstract:
Virginia Tech will integrate electronic theses and dissertations (ETDs) with large language models (LLMs) to improve digital library services for information discovery and research assistance.
Deliverables:
Integration of ETDs with LLMs for advanced information retrieval.
Prototyping software leveraging ETD-LLM tools.
Development of information discovery and research assistance tools.
Dissemination through publications, open-source software, and workshops.
Enhanced digital library services and tools for academic use.
Shoutout to my friends at Virginia Tech. The AI Hokies. Nice work on your two IMLS grants — happy to help if I can be of service.
