Libraries and AI:

AI is changing how the world works—and libraries are right in the middle of that shift. As trusted hubs for learning, libraries now can use AI not just to keep up, but to lead.

AI can help streamline operations, improve how people find information, and give staff more time to focus on what matters most—connecting with their communities. But as with any new technology, how we use AI matters. That means staying true to our values of respect, intellectual freedom, and user privacy.

Over the past two decades, libraries have evolved into dynamic, tech-forward spaces for discovery and lifelong learning. AI can support this shift through smarter systems for cataloging, better discovery tools, and more personalized services. This paper explores real-world ways AI is being used in libraries today—and how we can move forward responsibly, together.

 

AI tools using natural language processing (NLP) and machine learning are making it easier for users to find what they need—across languages, formats, and systems.

A great example of this is OCLC's AI-Generated Book Recommendations. OCLC has introduced AI-generated book recommendations in WorldCat.org. This feature uses artificial intelligence to suggest related titles based on user searches, enhancing the discovery experience without relying on personal data. Library Technology Guides

AI can automate metadata creation, improve classification, and surface links across collections, saving staff time and enhancing discovery.

In August 2023, OCLC implemented WorldCat De-duplication, a machine learning model, to detect and merge duplicate bibliographic records in WorldCat. This initiative led to the removal of approximately 5.4 million duplicates, improving data quality and consistency across the catalog. OCLC. The 2023 model focused on English language records and monographs. Lessons learned from that model have informed and updated OCLC’s current model, which can be applied against non-English and non-monograph records as well.

By analyzing usage data, AI can help libraries anticipate user needs, tailor recommendations, and reach underserved groups.

Wayne State University Libraries developed an AI application using TensorFlow to predict the likelihood of items being checked out based on their Library of Congress Classification Numbers and past circulation data. This tool aids in collection development and management decisions. OCLC

Chatbots and AI-powered assistants can handle routine questions 24/7, freeing staff for deeper engagement.

The University of Calgary Library implemented an AI-powered reference chatbot to support users with real-time answers to reference questions. Built on a Microsoft Azure framework, the bot is integrated with LibAnswers and trained on FAQs, enabling responsive and scalable support for patrons. Hanging Together Blog, 2024

AI models can guide acquisition, optimize shelving, and streamline digital storage based on use patterns—helping libraries do more with limited resources.

OCLC’s Smart Fulfillment uses real-time data to route interlibrary loan requests to the fastest lender, improving turnaround times and reducing workload. For example, the Big Ten Academic Alliance (BTAA) partnered with OCLC to implement smart fulfillment across its member libraries, significantly reducing average delivery times and improving user satisfaction. OCLC

 

AI must be used responsibly. Transparency, fairness, and accountability are essential. So is sustainability.

Responsible use means looking beyond what AI can do and asking whether it should. Every implementation should be grounded in real needs, not just novelty. Does the solution truly require AI, or could another approach be just as effective, with less complexity or environmental cost?

AI systems, especially large models, can carry a significant energy footprint. Libraries and their partners should consider environmental impact alongside cost, staff time, and community benefit. Balancing these factors helps ensure AI projects support the long-term mission of serving people thoughtfully and equitably.

Libraries must maintain their strong tradition of protecting user data with clear, ethical policies on AI usage.

OCLC's AI-generated book recommendations are designed with user privacy in mind, through a mandatory opt-in mechanism and prohibition on the use of personal search histories or data, aligning with libraries' commitment to protecting user information. OCLC Privacy Statement, 2023

AI changes workflows. Staff need support, training, and time to shift toward new roles in curation, digital engagement, and ethical oversight.

WebJunction offers an AI learning series to help library staff understand and adapt to the evolving landscape of AI in library services, ensuring they are equipped to manage new technologies effectively. WebJunction AI Panel

 

Artificial intelligence is not just a tool—it’s a turning point. For libraries, it offers a chance to reimagine how we serve, connect, and grow with our communities. Across discovery, operations, access, and engagement, AI is already making a real impact. But as we embrace its potential, we must do so with intention.

Libraries have always been leaders in responsible innovation. That leadership matters now more than ever. By aligning AI use with libraries’ core values and by supporting staff and patrons through education, we can ensure AI strengthens our mission, not shifts it.

The path forward isn’t one-size-fits-all. It will take experimentation, reflection, and collaboration across sectors. But with thoughtful planning, trusted partnerships, and a focus on people first, libraries can shape a future where technology deepens our impact—not just digitally, but humanely.

AI is here. Let’s make sure it works—for everyone.

 

Suggested citation: OCLC. 2025. “Libraries and AI: Managing Change, Advancing Access.” Dublin, OH: OCLC. https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17https://doi.org/10.25333/shzv-kk17.

© 2024 OCLC, Inc. All rights reserved. The following OCLC product, service, and business names are trademarks or service marks of OCLC, Inc.: Connexion, DDC, Dewey, Dewey Decimal Classification, Meridian, OCLC, WorldCat, WorldShare, and “Because what is known must be shared.” In addition, the WorldCat, WorldShare, and OCLC symbols are service marks of OCLC, Inc. Third-party product and service names are trademarks or service marks of their respective owners. OCLC grants permission to photocopy this publication as needed.