Sarah Calhoun and Sean Leahy, Reference Librarians, Carleton College

5 June 2025

Listen and subscribe to The Year of Curiosity: AI podcast on Apple Podcasts or Podbean.

Highlights from talk with Sarah Calhoun, Reference & Instruction Librarian for Humanities and Digital Scholarship, and Sean Leahy, Reference & Instruction Librarian for Social Sciences and Data, as they discuss research, reading, and what role generative AI may play in libraries.

Curiosity

George is curious about how colleges are using chatbots to engage potential students, after his 18-year-old encountered an aggressive chatbot on a college website. Would this lead to students feeling like the college is a friend before they arrive on campus?

Jennifer is curious about the origin of April Fool’s Day as the podcast was recorded on April 1, 2025. A web search gave results of unknown origin or largely unknowable, according to the Encyclopedia Britannica. Possible origins include pagan festivals or a reference to the 32nd day of March in Geoffrey Chaucer’s Canterbury Tales. The custom was firmly established by the 1500s, according to a 1561 Flemish poem about sending a servant on “fool’s errands” on April 1st.

vertically split image with Sarah Calhoun on the left and Sean Leahy on the right.

Sarah is wondering about the balance between innovation and preservation of government and the environment, with innovation often benefitting the few, while government and the environment are usually considered benefits for the many.

Sean is curious about the gift Reed Hastings gave to Bowdoin College to set up an AI initiative. What will this initiative look like, how will it develop, and is what we are experiencing right now a hype rather than reality?

AI Origin Story

Sarah has been aware of Ai since 2017, when she, together with Carleton’s Digital Scholarship interns, started researching AI: concerns about bias as it specifically relates to race and gender; AI use by police (Clearview AI); Google’s firing of its ethical AI team in December 2020, and the environmental costs of training AI.

Sean’s AI Origin Story is more recent. While he jokingly mentioned the Spielberg movie AI as a huge disappointment, for him, the technology really became a big deal in November 2022, with the release of ChatGPT-3. He realized that AI was still a focus after his parental leave.

Some Highlights

Sarah and Sean briefly outlined what a research librarian is and what their tasks are, with the recognition that their work constantly changes. Depending on the institution a librarian works at, the work will have a different focus, supporting deep research at larger universities versus focusing more on teaching undergraduate students at places like Carleton. They emphasized that their goal is to make information accessible, not hoard it, and they pushed back against the shushing librarian stereotype.

So far, the direct impact of AI on working with students has not been extensive, but it is changing. AI tools are becoming ubiquitous as databases and research platforms integrate them. Librarians anticipate their role will involve helping students navigate these different tools and teaching them how to think critically about them, which aligns with their traditional role in information literacy. They noted that friction in the research process can be valuable for learning and unexpected discoveries, expressing concern that AI might make things too easy and lead students to accept AI-generated content without critical assessment.

Tools like ChatGPT and Gemini may make the work for students harder, depending on the assignment. Students may struggle to assess what AI summaries include or exclude, and the tools can give the illusion of being a one-stop shop rather than one tool among many. A significant problem is that these tools can hallucinate, inventing citations or incorrectly combining information. Sarah shared an experience where Gemini gaslighted her about the meaning of an abbreviation in historical shipping news, sticking to an incorrect answer despite evidence to the contrary.

The impact of AI on scholarly publishing is an active area of discussion and concern. Publishers are seeking guidance, and there are ongoing court cases regarding the use of scholarly material to train AI models. Groups like COPE (Committee on Publication Ethics) are developing standards, and new licensing models are emerging, but there is a lack of specific limits. AI could lead to an increase in the sheer volume of publications, potentially exacerbating the problem of information overload while not really contributing to new knowledge.

Sean expressed worry about the threats AI poses to library labor, both direct (like the automation of tasks such as cataloging and indexing) and indirect (like a decrease in reference desk use potentially being used to justify cost-cutting or rightsizing libraries). Large university systems are investing in AI partnerships that include access to powerful reasoning models which might mimic the work librarians do, potentially making human librarians seem expensive and redundant. Sarah added concerns about AI impacting fundamental concepts of reality and truth in teaching. The library is actively involved in tracking AI developments and relevant legal cases.

Both agreed that if one could offload anything reliably to AI, doing taxes, shoveling sidewalks, and meal planning would be amazing ways to use AI.

Tools/Recommendations

Return to AI at Carleton Blog