When AI produces a convincing answer, how do you know it is right?

Multi LLM Hackathon for Strategic Intelligence in Education & Training

Most people use one AI tool at a time, for one task at a time. A quick question to ChatGPT, a summary from Claude, a draft from Gemini. What almost nobody does is put several of these tools to work together, deliberately, on a single piece of research, and hold the result to the kind of standard a decision-maker could actually rely on.

That is the premise behind this full-day, pre-conference workshop from Paul Bacsich, Margaret Korosec and Glenda Morgan. They call it an AI Evidence Sprint. There is no code involved, and nobody is being taught how a language model works. The question is sharper and more practical: can a small team, using the frontier AI tools they already have access to, produce a piece of strategic research that is genuinely trustworthy, inside a single day, on a question that matters to their own institution?

Anyone who has asked a chatbot for a research summary knows the catch. The output reads confidently, the structure looks plausible, and somewhere in the middle sits a claim that turns out to be invented, misattributed, or simply wrong. That gap, between how AI-generated research sounds and how reliable it actually is, is exactly what this workshop is built to close.

Working in small groups with their own LLM subscriptions, participants take a real strategic question in education and training and turn it into a structured, referenced, independently verified briefing by the end of the day. It is a workshop measured not by how much text a group can generate, but by how much of that text they can stand behind.

About the facilitators

Paul Bacsich has worked in online learning and e-learning consultancy for close to four decades. He is CEO of Dualversity, founding director of Matic Media Ltd, and a Senior Associate at Sero Consulting’s SeroHE division. His long career spans the UK Open University, a professorship in Telematics at Sheffield Hallam University, and consultancy for universities, governments and international agencies including the European Commission and UNESCO. His current work looks at how large language models are reshaping assessment practice in higher education.

Margaret Korosec is Dean of Online and Digital Education at the University of Leeds, where she leads the Digital Education Service and founded Helix, a space for experimentation in digital learning. She previously led the growth of the online portfolio at the University of Derby Online Learning, and began her career in online and competency-based education at Western Governors University in the US. She holds a PhD from the University of Hull and is a Senior Fellow of the Higher Education Academy.


Brian Mulligan is a leading expert in learning technologies, with a career spanning three decades in shaping digital education in Ireland. As the founder of the Irish Learning Technology Association in 2002, he has played a pivotal role in advancing the field. At Atlantic Technological University, he spearheaded the development of online distance learning, driving innovation in accessible education. Now a consultant with Universal Learning Systems (ulsystems.com), Brian specialises in the strategic application of learning technologies to transform education and expand opportunities for learners worldwide.

Full details on the workshop and booking are available here. Seats are limited.

Written for OEB 2026.