This sounds deceptively simple. Information professionals are human, thus being in the loop should come naturally. But it does raise questions about which humans should be in the loop, what loop are we talking about, and where knowledge transfer comes into the equation. Plus, way too many people are putting total trust into results they get from querying AI chatbots without stopping to think about validity.
Keynote speakers at Knowledge Summit Dublin, held on the Trinity College campus 29-30 June 2026, addressed various elements of the convergence of AI, knowledge and humans. Starting off was Sean Lally, Founder, The Lean Learning Professional, who stressed the importance of struggle, which he defined as the most important part of learning. AI does not struggle. It can improve performance but it does not create durable learning. Although a summary can feel like understanding a topic, that is a fluency illusion. For true learning, we should embrace the struggle.
Gianni Giacomelli, Head of Design Innovation, MIT CCI, took Lally’s concepts a bit further as he considered the new AI frontier. Although humans and machines are social learners, they differ in how information is processed. He sees a progression from sensing to remembering to creating alternatives to deciding to act and, finally, to learn. Our ultimate goal, he said, is to steer the behavior of the AI ecosystem.
The Unlearning Lady, Zanela Njapha, brought the concept, not only of learning, but more importantly, of unlearning and relearning. Outdated mental models work against us, as does peer pressure. As technology advances, our ingrained approaches to our jobs need to change and adapt. Listening is an important skill in our new environment so that we can lead with agility and lose our sense of certainty. Watch out for the words "clearly" and "obviously", since the reality is that it probably is neither clear nor obvious. We need to challenge our assumptions about what we know and focus on relearning.
Learning by Example
Breakout sessions during Knowledge Summit Dublin touched on issues important to internet librarians, such as neurodiversity, futurism and achieving buy-in for projects. The conference operates on a flipped format for the breakout sessions: 20 minutes of presentation followed by discussion and/or an exercise. Some suggestions for making the workplace better for neurodiverse employees were to share slides and agendas before any training sessions, offer written alternatives alongside verbal, make things visible without requiring people to perform visibly, and build in reflection time.
Contemplating the future had groups given three cards with possible future scenarios and asked us to brainstorm what that would mean for the workplace and for society. At my table, they were "We all communicate in verse", "Oversimplification leads to fragility and bitterness" and "Body modifications challenge what it means to be human". Can you imagine getting all reference requests in verse? Not sure I can and I admit I’m still contemplating that last one about body modifications.
Emory Consulting’s Rachel Teague and I presented on persuasion techniques to get buy-in for projects. Most of my portion came directly from my experiences as a corporate librarian, including tailoring your language to that of the decision maker, making the value of what you want buy-in for impossible to ignore, explicitly stating the cost of inaction, starting with a story and aligning it with their goals not yours.
Saving Knowledge
Exciting was the launch of the USAID Knowledge Rescue Initiative during the conference. An all-volunteer project, it has created a platform to restore the 60 years of information and documentation at risk of being destroyed after the agency was summarily closed in February 2025. The initial tranche consists of 112,000 files, with more to come, along with hundreds of interviews and oral histories from former USAID staff. The Initiative involved collaboration among communities from over 100 countries and is making data freely available to practitioners, researchers, and organizations working to create positive change. Equally exciting was the art auction during Knowledge Summit 2026 that raised over 800 Euros for the Initiative.
When it comes to data rescue, volunteering, and fund raising, humans are definitely in the loop—and many of those humans are librarians and other information professionals. When it comes to AI, however, who should be in the loop is not quite so clear. Looking back at lessons learned from Knowledge Summit 2026, I see a shift in our role. It’s no longer simply reviewing AI-generated content, challenging results emanating from prompts, discerning when AI reasoning went wrong, and being accountable for providing quality information. It’s expanded into overseer, curator, referee and guardian of truth. As agentic AI becomes commonplace, our role in guiding its activities becomes stronger. Do we step up to the plate for this or do we leave it to IT?
Librarians are good at communicating with our users and at connecting people with information. AI is not going to replace humans but it is going to change the nature of our work. Human judgement becomes more important, essential even, as AI gains ground. Autonomous decision making by agents needs to be governed by humans. Being in the loop understates our role since it’s a bit passive. Plus, library environments have many loops and information professionals should have a presence in them. Our role goes beyond those loops; it's about being the connectors between humans and technology and between human and human. We should be active collaborators with generative and agentic AI, not to mention whatever AI comes next.