AI at ACRL 2025: Theory, Applications, and Debate 

Reflections from the ACRL conference

Posted on April 9, 2025 in Blog Posts

Many have come to know LibTech Insights for our great posts on user experience, digital preservation, and pedagogy, but we like to think of ourselves as the premier party reporters in the library technology space. ALA Annual? Got you covered. Charleston? Been there. ITHAKA? We were the life of the party. So, of course, we weren’t going to miss the ACRL 2025 conference in Minneapolis. We will get into the highlights in just a sec, but right off the bat, I want to let you know that the event’s signature mocktail was a saccharine, pink-Starburst-flavored drink and that spring/summer’s hottest item, bucket hats, seems slow to hit the academic librarian community (“buyer beware” or “fortune favors the bold”?). 

Although I intended to offer more panoramic coverage of the library tech conversation at ACRL, I found that the conference tilted heavily toward AI. Indeed, a solid eighth of the panels centered on AI, which is to say nothing of posters, roundtables, and other presentations. This was understandable: AI is a hot topic, particularly within ACRL, and a conference is a good place to engage in high-level discussion. I hope to offer a limited survey of what I heard and saw during my time at ACRL. 

Theory: How should we train ourselves and others for an AI-driven information environment? 

Librarians have widely acknowledged that AI intersects with ongoing professional concerns for information literacy, but the precise relationship between AI and information literacy remains a hot topic.  

Taking full advantage of the conference, ACRL’s committee on AI literacy held an hour-long discussion on its draft of AI Competencies for Library Workers (based on Sandy Hervieux and Amanda Wheatley’s fantastic white paper for Choice). ACRL President Leo Lo made the case for AI training for library workers, particularly in academia, which has been one of the first impact zones of AI. Lo held that AI literacy is a natural extension of information literacy because AI “impacts all aspects of information discovery, use, and evaluation.” Leadership in this area will allow librarians to “proactively shap[e] our role rather than reac[t] to the changes.” Understanding that librarians hold a variety of feelings about AI, Lo argued for a neutral and critical approach, which is neither pro- nor anti-AI but “pro-learning about AI” while still grounded in the ethical values of librarianship. 

I won’t get into the competencies themselves—I encourage you to read the draft document—but I wondered if we couldn’t learn something from J. Kevin Sebastian’s critique of AI literacy from his talk “Reframing Information-Seeking in the Age of Generative AI.” Sebastian argued that an approach to information-seeking guided by the vocabulary of queer theory might better conceptualize this new information environment. Traditional scholarship on information-seeking behavior, Sebastian noted, emphasizes efficient information retrieval: people have a question, go online, and find an answer. But AI is different because it isn’t just retrieving information—it’s also synthesizing it.  

For Sebastian, a queer theory approach sees knowledge not as an end point but as a construction. Hence, in his words, “A queer AI literacy disrupts stability, embraces fluidity, and centers embodied experience. Generative AI isn’t just reflecting knowledge—it’s performing it. And that performance is always situated, partial, and political.” AI literacy should move beyond seeing AI as a tool to treating it as a collaborator. 

Certainly, there are pros and cons to these approaches. AI models may act as collaborators, but they are also products, and thinking of them as such, as someone raised in the Q&A portion of the ACRL AI task force panel, offers librarians leverage in shaping AI in their interactions with vendors. But in both cases, I think what’s important is that librarians feel empowered to negotiate the terms of engagement with AI. 

---

🔥 Stay up-to-date with LibTech Insights by signing up for our free newsletter. Just one weekly email with our new blog posts, top tech news stories, and other bonus content. Check out some posts from our archive:

• LTI’s Complete Toolkit for Preserving and Restoring Web Pages
• The Next Chapter: Academic Libraries Enter the Intelligence Age 
• Using Notion to Organize Library Projects

---

Applications: How can we implement and evaluate AI? 

Though these debates over theory and training are interesting, I think many librarians just want to know whether AI can actually help them with their jobs. (If this is you, sign up for our free webinar on using free or low-cost AI tools to enhance library workflows.) Kerry Lohmeier, Beth Jennings, and Suzanne Darais offered a fantastic panel highlighting some uses they have found for AI. Here are some of my favorite suggestions: 

• If you like a feature from a website you’d like to include on your own, screenshot it and ask ChatGPT to mimic the code. You can also add restrictions, e.g., “I don’t have access to the stylesheet, so keep your response to HTML only.” 

• For collection development librarians, use AI to create a starting point for weeding projects, e.g., “Weed books under subject heading B that are older than 15 years AND circulated 0 times in the last 7 years EXCEPT for core philosophy texts, major religious texts, or interdisciplinary works related to X.”

• Have it improve the design of emails, e.g., “Format this email to make it more likely that people who don’t usually respond to emails will RSVP to this event.” 

The Choice team also discussed the rubric we developed for evaluating AI products and tools. Librarians Abby Mann (Illinois Wesleyan Univ.) and Evan Fruehauf (Univ. of South Florida) offered case studies from a small and a large library, respectively, to discuss their own concerns when evaluating AI products. As Mann observed, small libraries have scarce time, resources, and staff, so they have to be strategic about which tools to consider deploying. (See Mann’s LibGuides.) Fruehauf said his university has become very pro-AI, but this hasn’t eliminated the learning curve with AI tools or more general concerns with AI, e.g., privacy. (See Fruehauf’s LibGuide.) 

Debate: Should librarians engage with AI? 

Humming beneath the surface of many panels was a quiet debate. Should librarians endorse, recommend, or otherwise engage with AI technologies given the large ethical issues (from environmental destruction to copyright infringement to misinformation)? In the feedback shared from ACRL’s draft competencies for AI literacy, many respondents wanted the competencies to reflect a decision to “opt-out” of AI or saw the competencies as too “pro-AI.” “Knowledge of AI tools is essential,” an anonymous librarian noted in their feedback. “Use is not.” 

To my knowledge, none of the panels at ACRL were openly anti-AI, which is disappointing; I think the perspective is worth hearing. But many of the “pro-AI” panels gestured toward this debate or the speakers’ position in it. For instance, Sebastian contended that refusal to engage with AI by claiming the moral high ground might be myopic toward the needs of the people librarians claim to serve. Forcefully, he argued, by refusing to engage, “[w]e forfeit our ability to shape [the system].” 

That said, as I raised in Choice’s panel on AI, I do wonder whether anti-/pro-AI is the best framing for this debate. My hunch is that these positions represent the ends of a bell curve, ignoring the vast number of people in its middle who haven’t quite made up their minds. After all, what it means to be “pro” or “anti” is largely, and perhaps intentionally, left undefined. Does “anti-AI” mean totally opting out (to the extent one even can)? Does “pro-AI” mean anything other than total refusal? At times, it seemed this way. 

However, I would posit that the spectrum between skepticism and curiosity better captures most opinions on AI. Many of the AI panels I attended were overflowing with people who were interested in learning what’s out there and evaluating it for themselves. I worry that the pro/anti division will shut down or warp this conversation. None of us were responsible for opening Pandora’s box, and none of us can close it. The question of what we do now is one that requires active engagement and intellectual generosity. As we saw at the conference, that’s surely something we can give to one another. 

Related Posts