ASIS&T president Ian Ruthven welcomed attendees, saying, “AI is the defining technology of our generation.” Speaking after Ruthven, incoming president, Maria Bonn, referenced the event’s theme by stating that an ASIS&T conference is the easiest place to have difficult conversations. Both proved correct.

The conference oscillated between presentations describing research projects and more free-flowing panel sessions. The research presentations will appear in the conference proceedings, which are in the ASIS&T Digital Library, published by Wiley.

Agentic AI 

John Symons, professor in the department of philosophy at the University of Kansas and director of the Center for Cyber Social Dynamics, gave the opening keynote address, Artificially Intelligent Agents, Domain Specificity and Inquiry. How to bridge the gap between structured research and messy social values is on his radar. Agentic AI, he thinks, promises a partnership in inquiry rather than a supplement. As a partner, it can automate discovery and generate hypotheses. There are problems, however, with improper use of training data and lack of reproducibility. Plus, no algorithm exists that outperforms every other one in each and every scenario. You need human-centered AI, he believes.

Symons presented several guidelines for scientific discovery. Agents perform best when the task is extremely narrow. Use gamification, and ground the agent in physical reality. Force the agent to judge itself. Train models on messy data. Don’t discount the role of serendipity, and recognize that AI can distort the role of imagination, which is something humans excel at.

Going well beyond the caveats expressed by Symons, a session on AI in libraries centered on misinformation, human-centered AI, and reference services. Oghenere (Gabriel) Salubi noted that AI accelerates information disorder and gave real-world examples of fake news, identity fraud, and political propaganda. He also said that bots amplify disinformation, and disinformation circulates faster than corrections. Ian Song cited the European Union’s passage of its AI Act in 2024 as indicative of how AI technologies should be tools that serve people. Jose Aguinaga presented on the evolution of reference services, which includes personalized reference services, enhanced community-building, conversational assistants, and ethics. Issues exist with hallucinations, governance, and regulation. Another issue was raised by Hannah Moutran in her talk on archive metadata and transcription: When applying AI tools to historical content, do models transcribe words now considered harmful and offensive as written, or do they flag for human review?

Other talks dealt with agentic AI as well. How generative AI changes the work, roles, and priorities of libraries and information professionals was addressed in a spirited discussion. Student reactions to AI vary. Some are afraid to use it, fearing that would be perceived as cheating and they could be expelled. In other organizations, students are encouraged to use Microsoft Copilot since the university has a campus-wide subscription. Training in prompt engineering would help, as would meeting students where they are to aid them in using AI. A balance is necessary between productivity and over-reliance on AI. Downsides of AI, including the environmental impact of data centers and the dependence on commercial entities to build library products, raise many questions. 

Measuring the knowledge breadth and depth of large language models (LLMs) attracted the attention of Xueying Peng, Quan Lu, and Kehui Lui. Their systematic review and integrated framework development looked at how generation capability and knowledge management tasks matched. The framework ensures diversity and avoids redundancy. LLMs were also examined in light of how well they could automatically generate abstracts. Yumi Kim, Jongwook Lee, and Seungwon Yang found that experts actually preferred the AI-generated ones over author abstracts.

Closing out the November conference was a debate with this declaration: “The information professions must play a key role in the development, implementation, and use of AI.” It may come as no surprise that the team in favor of the resolution prevailed. The “For” team comprised Lynn Silipigni Connaway, Souvick Ghosh, and Yuan Li. “Against” was Gary Marchionini, Isto Huvila, and Dania Bilal.

Information Behavior 

The December virtual satellite meeting began with the President’s Lecture, presented by Ina Fourie, emeritus professor of information science at the University of Pretoria in South Africa. In her presentation, Information Behaviour as a Research Lens for Agility in Change Enablement, Transition, and Flow—Workplace and Everyday Life, Fourie stressed the importance of information access and delivery to improve people’s quality of life. “We no longer have boundaries and need to find balance between work and personal life,” she said. 

The rapid developments in IT, AI, and other aspects of life engender uncertainty that calls for a shift in practices to overcome information barriers, according to Fourie. “Whole brain thinking” affects how people approach change. Change management focuses on organizational systems, timeline, procedures, and processes, but context is important, as is the recognition that it is human nature to resist change, and limitations in habits, priorities, and mechanisms for monitoring change are potent. Fourie remarked that information behavior is emotional, behavioral, cognitive, adaptive, and practical. Ultimately, agility and change-enablement must honor human dignity, emotional well-being, and the rich complexity of how people live, learn, and adapt. Her closing thought was, “When considering the turmoil and challenges of contemporary society and the need for agility, a much deeper look, inspired by interdisciplinary insights and self-awareness, is required.” Information behavior is the research lens to enable understanding.

SLA at ASIS&T 

What wasn’t a difficult conversation was that between SLA members and ASIS&T members. This was the first conference after the merger of the two associations. Although few SLAers attended the conference, ASIS&T organized a reception for the local D.C.-Maryland-Virginia SLA chapter one evening prior to the conference, which was well-attended.

The ASIS&T conference had a more academic vibe than that of traditional SLA conferences, which tend more toward the practical. But have no fear; ASIS&T has a plan to make SLAers feel at home in the association. The Information Science Summit and SLA Conference is scheduled for June 6–9, 2026, in Albuquerque, New Mexico, with the theme Exploring the Frontiers in Information Science and Librarianship. It promises to bring together librarians facing “wicked information challenges” in their jobs with those who study how to solve those types of challenges. That conference is likely to be the test as to how well the former SLA can be integrated into ASIS&T.