CARL at 50: A History of AI and Libraries
October 14, 2026 – This special 50th anniversary edition of CARL’s AI Digests was prepared by Amanda Wheatley, Associate Librarian at McGill University and CARL’s Visiting Program Officer (VPO) for AI and Library Strategies.
The term artificial intelligence was officially coined in 1956 at the Dartmouth Summer Research Project on Artificial Intelligence. Two decades later, in the mid-1970s, libraries began to feature in the scholarly literature around AI and automation. As CARL celebrates its 50th anniversary, this digest considers some of the first papers on AI and libraries, one per decade from 1976 to 2026, to demonstrate just how far we have come.
“If a computer could diagnose bacterial infections and prescribe treatment why couldn’t a computer diagnose an information need and present a reference book to solve the problem?”1.
In 1976 – the year CARL was established – Linda C. Smith argues in “Artificial intelligence in information retrieval systems”1 that AI can make information retrieval systems more capable by improving how they represent problems, interpret language, and respond to users’ information needs. This once revolutionary concept now seems easily achievable when reflecting on current use of AI technologies in these systems. Linda Smith’s research resonates now because her arguments go beyond simple search and retrieval of documents, but look at how AI might be used to retrieve answers to questions. Her work 50 years ago presents a vision of AI-powered information retrieval that feels modern.
Jumping forward to 1986, Karen Smith’s “Robot at the reference desk?”2 explores a practical application of using AI expert systems to suggest readings based on a computer-powered reference interview. She asserts that people don’t want to read the descriptions of multiple different reference books to get to the right one, they want individualized attention. She worked with computer engineers to create Pointer, a system designed to retrieve sources related to government information. Interestingly, her endeavours align closely with recent efforts by librarians to create their own expert systems and large language models to interact with their collections.
Advancing another decade, the theme of AI use in the library moves toward human-guided systems, as seen in B.A. Nardi and V.O. Day’s 1996 paper, “Intelligent agents: What we learned at the library”.3 Here, Nardi and Day, two researchers from the technology industry, study librarians as a model for how intelligent agents should behave. Their research found that librarians were adept at handling personalized research queries and that software agents should move away from one-size-fits-all systems toward richer representations of users and context. One of their suggestions was to have librarians co-design AI client agents and theorized this could be a new role for librarians in the future.
In the new millennium, the research around AI in libraries begins to quiet. What was once a lively area of scholarship, slows down, though there are still some articles through which to track the technological progress of the decade. In 2006, Lili Li’s paper “Emerging Technologies and Their Impacts on Academic Libraries, Librarians, and Library Information Services”4 gave space to an AI system designed to interpret human learning for education. The paper looks at the company Blackboard Inc. and their technology for electronic whiteboards that can be used for interactive teaching. Though this exploration into AI is brief, it shows that librarians are still interested in the field.
In the 2010s, the lull in scholarship related to AI and libraries persists, though a ramp up is imminent. In 2016, Peter Fernandez’s Library Hi Tech News series “Through the looking glass: envisioning new library technologies” and the two-part series on AI provides a glimpse into the sentiments of the time. Fernandez’s first article “Understanding artificial intelligence”,5 is primarily concerned with reintroducing AIto librarians and establishing a new baseline of foundation knowledge. In his second column, “How artificial intelligence will impact libraries”,6 Fernandez takes a closer look at the potential for AI to impact libraries due to increasing complexity of the technology.
Today, in 2026, research surrounding AI and libraries is abundant. Coverage includes wide-ranging topics such as machine learning for metadata extraction and cataloguing, textual and visual analysis, information seeking and retrieval, AI literacy, and more. Looking back fifty years, however, it is clear that librarians have always been curious about the potential impacts of AI.
References
- Smith, L. C. (1976). Artificial intelligence in information retrieval systems. Information Processing and Management, 12(3), 189–222. https://doi.org/10.1016/0306-4573(76)90005-4
- Smith, K. F. (1986). Robot at the Reference Desk? College & Research Libraries. https://doi.org/10.5860/crl_47_05_486
- Nardi, B. A., & DAY, V. O. (1996). Intelligent Agents: What We Learned at the Library. Libri, 46(2), 59.
- Li, L. (2006). Emerging Technologies and Their Impacts on Academic Libraries, Librarians, and Library Information Services. International Journal of Technology, Knowledge and Society, 2(4), 109-122. https://doi.org/10.18848/1832-3669/CGP/v02i04/55570
- Fernandez, P. (2016). “Through the looking glass: envisioning new library technologies” understanding artificial intelligence. Library Hi Tech News, 33(3), 20-23. https://doi.org/10.1108/LHTN-03-2016-0013
- Fernandez, P. (2016). “Through the looking glass: envisioning new library technologies” how artificial intelligence will impact libraries. Library Hi Tech News, 33(5), 5-8. https://doi.org/10.1108/LHTN-05-2016-0024
