“It’s Not Gospel, It’s One Nugget of a Wider Process”: How Researchers Question, Verify, and Trust LLM Outputs

Academic researchers face challenges when verifying information generated by LLM-powered tools during information-seeking tasks.
This talk investigates the factors that make verification difficult and the strategies researchers adopt to overcome these challenges. It examines key characteristics of current LLM-based tools and explores how we can design systems that researchers can trust.

About the speaker:

Monica is a PhD student at the Human-Computer Interaction Design (HCID) Centre at City St George’s, University of London, where she researches misinformation in the use of Generative AI. Originally trained as an Industrial Engineer, she later moved into Information Technology, working as an analyst/programmer, project manager, business analyst, and product manager. Her interest in user-centred design led her to pursue a Master’s degree in Human-Computer Interaction with Ergonomics (HCI-E) at UCL. Her current research reflects a longstanding interest in understanding how people make sense of information, technology, and their lived experiences.