Computers in Libraries was held March 17–19, 2026, just outside of Washington, D.C. I was able to make it to two of the keynotes and a few sessions from the various conference tracks on AI, leadership, and access and discovery.

AI CHANGING RESEARCH

Dan Russell, a “free range research scientist” who worked at Google and has taught at multiple universities, gave the keynote session AI, Search, and the Future for Information-Finding Experiences. Russell calls himself a cyber-tribal-techno-cognitive-anthropologist because he “think[s] about how groups of people think about, understand, and use the technology of AI systems.” The best researcher of the future will have a different skill set from the researcher of today, he believes. They will work with AI to advance their field of study. Currently, AI is what Russell called co-intelligence—AI is like an intern that you have to explain your research to because they can’t work with keywords alone.

Russell shared some real-life AI examples, such as finding specific topics in his Gmail messages and analyzing his exercise routine to gain insights on the areas that need improvement. AI offers the automation of repetitive tasks, he said, as well as content creation and editing help, enhanced communication, and smarter virtual assistants. Ask yourself what you can do with it to improve your research, he suggested. Another example he provided was offering an AI tool a photo of his disorganized bookshelves; he asked the tool to find a specific title, and it replied with the exact location, even though the book was partially obscured beneath another title. While sharing even more examples of how he uses AI (language translations, creating an interactive periodic table, etc.), he cautioned that AI output always needs to be fact-checked. An important skill to learn when working with AI is the ability to give it clear directions. He likened our age of AI to an age of centaurs—i.e., a human head is driving an AI body. Human judgment must remain central to AI use, he said.

Russell made predictions for the future of AI, which include:

• There will continue to be hallucinations resulting from a lack of humans checking outputs.
• Internal AI bots in organizations will become common for managing data.
• Idea development will speed up.
• Research skills will atrophy. (Russell said he is already forgetting how to code manually.) 

His advice for future-proofing is to engage with the tools so you can start learning, because anyone can look something up, but it takes initiative to take action. Cultivate your intuition by practicing correct decision making amid speedy executions, he added, and stay curious by keeping track of the latest developments.

Students need guidance on AI, which is how libraries can help. Russell noted that libraries’ role is “[t]o help researchers and patrons by creating the best possible content, providing the organization, instilling the motivation, [and] making the environment one that leads to inspired learning.” Russell called research a socio-technical skill because it encompasses engagement with both people and technology. The tools are extraordinary, but, he emphasized, they need human judgment to be their most effective. 

BOTS RULING THE INTERNET

Lee Rainie, director of the Imagining the Digital Future Center at Elon University, presented the keynote session Humans, AI, Bots, and Engaging With Information. He opened his talk by describing some of the recent news stories about AI, including the Pentagon’s blacklisting of Anthropic, which he said “makes the fog of war foggier,” and McKinsey & Co. stating that 25,000 of its 60,000 employees are AI agents. Rainie then discussed Moltbook, a social network exclusive to AI agents, which baffled the crowd. It gets even weirder—Rainie pulled up the homepage of the Church of Molt, which is a new religion the Moltbook-using AI agents created themselves. The human instinct to be social is part of AI agents’ training, Rainie explained. He discussed AI adoption habits among U.S. adults, noting that for a variety of metrics, majorities say they use AI agents. This popularity sits alongside existential concerns, Rainie clarified—people use AI, but they tend to be wary of it.

Bots are close to dominating the internet in that their output will soon overwhelm human content on the web, Rainie noted. This is “disastrous” to aspects of the web such as referral traffic to websites from Google searches and how much AI steals others’ content. He talked about the rise in teen use of AI and how social use is increasing; for example, AI as a dating coach. There are other ways bots are starting to seem human, and Rainie shared some instances: They can tell when they’re being tested or a joke is being played on them, and they can play tactical games. Some can even override instructions to shut themselves down.

Libraries are in the thick of grand questions about AI, Rainie continued. We’ve never faced an intelligence greater than human intelligence until now, he said, with the tools challenging our cognitive capacity. Human agency is at stake, so libraries’ call to action is to encourage humanness. The following are some suggestions he provided for libraries: 

• Serve as “experience orchestrators” of learning and engagement with media.
• Become community ethnographers. Develop digital archives of oral histories, community narratives, and underrepresented voices to feed more inclusive AI training sets.
• Use AI to augment digital preservation (e.g., repairing old documents, organizing historical material) without losing human curatorial control.
• Partner with artists and scholars to document and reflect on society’s evolving relationship with AI.

YOUNG PATRONS AND AI

Nick Tanzi, assistant director of South Huntington Public Library in New York, presented Future-Ready Youth: Preparing Communities to Navigate the Age of AI, which shared details about how young people in particular are using AI and how libraries can help them harness it in positive ways. He noted that AI is omnipresent on the internet; children and teens are “awash in AI output” that comes to them both in direct and indirect ways via their phones, social media, smart speakers, and more. Even the algorithms on YouTube Kids leverage AI. Studies show that majorities of students use AI—not to cheat at school, but for personal reasons, such as dealing with interpersonal conflict or their mental health. Tanzi called this “high-stakes usage.”

Caregivers tend to be unaware of how their children are using AI, which is a gap we need to bridge, Tanzi believes. He is ambivalent about AI; it has potential to be both beneficial and harmful, and he wants to mitigate the harm. Like the loss of social skills children had to face during the pandemic, AI could cause a lack of etiquette. AI could also have a negative impact on critical thinking skills and cause cognitive offloading—i.e., bypassing the learning process instead of engaging in it. Tanzi spoke about the headlines sharing how AI chatbots have encouraged young people to self-harm and to die by suicide. Libraries need to make their communities aware of these dangers, he said.

Libraries can help by promoting AI literacy and talking about how ethics and environmental impacts are tied to AI. Tanzi shared his most important areas of focus when it comes to AI education: privacy and internet safety, information retrieval and evaluating sources, and demystifying AI tech. This last focus would involve offering patrons the chance to experiment with AI tools, but that may require young people to get a permission slip signed. Libraries could avoid that by showing how to evaluate AI outputs. Tanzi stressed that no matter a library’s approach, it should teach both the utility of the tools and the ethics issues surrounding AI.

Tanzi shared some examples of AI education programming. An AI Sleuth contest involves putting an AI-generated image on a poster and asking patrons to list the reasons they believe it to be an AI image, then choosing a winner. The library can post the answer key next to the image for future patrons to see. In a Two Truths and AI game, a library could display posters of three images and ask patrons to identify which one of the three is AI-generated. This doesn’t feel like homework, Tanzi noted, but they’re still learning about AI. Any AI programming involving young patrons needs to be preceded by conversations with key partners such as parents, the local school district, and local colleges to empower them to talk to young people about AI. In conclusion, Tanzi emphasized that libraries should continue to be a place of human interaction and human-based services, where creativity is supported, critical thinking is encouraged, discovery is facilitated outside of algorithms, and being a safe, accepting space is prioritized. 

CHOOSING AN INSTITUTIONAL REPOSITORY

Metadata, Middleware, Mission: The Institutional Repository Journey at Howard University featured Howard University’s Kimberly Prosper, STEM librarian, and Qaddafi Sabree, emerging technologies librarian, who walked attendees through their process of creating an institutional repository. Prosper said they realized Howard could use one when it became clear that faculty research was scattered across the internet. A digital platform existed for the university, but it wasn’t designed for active faculty output; it was more useful for preserving historical collections. So the library set out to find a platform that encompassed the full scope of what it would need, including the ability to track impact metrics, a space for students’ theses and dissertations, and room to grow.

The library created a checklist that would enable the evaluation of each vendor it spoke to. The formal selection process had five components: 1) the wish list, which defined what the library would need in a repository platform before any vendor got involved; 2) proposal invitations—i.e., formal requests to all vendors under consideration; 3) live demos that each vendor showed the library so it could see the platform in action; 4) testing, which allowed the library to try out each platform and get hands-on experience with it; and 5) a partnership with IT, which checked that each platform met that department’s standards.

Sabree stressed the importance of involving the IT department early in the process. The library’s non-negotiables included integration with Howard’s existing systems, flexible storage, analytics reporting, ease of discovery, accessibility features, and a low-cost subscription. In addition, the library wanted to be able to include datasets, multimedia, archival materials, and the rest of the full spectrum of the university’s output in the repository, he said.

One of the challenges the library faced was faculty adoption—faculty members had to be convinced to agree to include their work in the repository. The library made sure they knew the benefits: They could use it to track their research impact (i.e., citations), collaborate with colleagues, and get permanent URLs for their work. An institutional repository also increases discoverability and contributes to more prestige for the university. Sabree noted that although the library has a metadata librarian, there needed to be a dedicated metadata lead for the repository who would adhere to a metadata policy that would ensure consistency. He is hoping someone will be hired soon. Another realization was that the users are both faculty members and students with their own needs. What they have in common is the desire for an easy submission process to the repository and tools to be able to track citations and share their work. The stakeholders have differing needs too, in that the administration wants a cheap platform, and IT wants a secure platform.

Sabree shared his list of what a “good” institutional repository looks like:

• Faculty members deposit their work into it without being asked.
• Metadata is consistent across collections.
• Downloads and citations get reported to institutional leadership.
• IT manages the repository like any of its other systems.
• New content types are added proactively so the repository can evolve with research practices.
• Students start mentioning the repository in their works’ acknowledgments so the word can spread to future researchers.

Sabree described the results of their platform tests; the winner was Digital Commons from Elsevier because it was the best match for them. In closing, he shared his lessons learned, which include broadening the vendor pool, running a pilot phase only for faculty members, and planning how to do metadata management earlier. Then he discussed the plans for the launch of the repository, which is scheduled for sometime in the next 6–12 months.

HOW TO CRAFT AN EFFECTIVE PITCH

Get Out of the Library: Influencing/Pitching Ideas was presented by Maurice Coleman, principal of Coleman & Associates, and M.J. D’Elia, CEO of Thirdway Think and director of the LLEAD Institute. Coleman immediately encouraged participants to gather in groups around the tables at the session to discuss what excites them about their job. After a few minutes, he explained that pitching ideas is about talking to people you don’t usually talk to, so he wanted the participants to do just that. The most important aspect of selling people on libraries, he said, is providing the WIME: “What’s in it for me?” 

D’Elia spoke about his background in sales, noting that pitching ideas is all about relationships. He discussed the six sources of influence that move people toward specific behavior (see Table 1) and the six actions necessary to move them toward it (see Table 2).

Table 1

Table 2

D’Elia said you need to engage with at least four of the six action boxes from Table 2 in order to move the needle on a pitch. He asked participants to get back into their breakout groups so they could brainstorm what accomplishing the four would look like at their library. Coleman and D’Elia walked around to provide guidance when needed.

D’Elia then explained the “seven simple moves” librarians can make to create an effective pitch. They should do the following:

1. Have a clear, simple description of an idea to present to a potential partner.
2. Describe why the initiative you’re proposing should exist. (It’s meeting a challenge, filling a need, seizing an opportunity, etc.)
3. Create a direct statement on why it matters to your audience.
4. Describe what you can accomplish together to benefit the community.
5. Describe how you will make participating easy for them.
6. Name a single, clear action you’re asking them to take.
7. Name a concrete starting point—a first step—to begin the collaboration.

To conclude the session, Coleman and D’Elia asked participants to speak in their breakout groups one more time to test their pitches. One of the librarians in my group was ready with a pitch using the seven steps, and the rest of us provided suggestions and feedback. Coleman reminded the participants that it’s important to give people a reason to care about your pitch; it helps prove your worth.