From Trust to Practice: AI and the Clinician Experience
AI is no longer confined to pilots and proof-of-concepts. It is sitting inside clinical workflows, informing decisions, flagging risk, and reshaping what clinicians do day-to-day. Technical accuracy tells only part of the story. Clinician trust, shifting accountability, and changes to the clinical role matter just as much.
This event looks at the behavioural realities of AI adoption in clinical practice: how a system lands in a ward, a clinic, or a consultation room. As AI tools move further into everyday healthcare, trust, deskilling, and implementation shape whether clinicians use these systems, and if patients are better off when they do.
We are bringing together academics, computer and data scientists, clinicians, and regulators to explore what it takes to move from a working model to a trusted, adopted tool in real clinical settings, and where behavioural science actually fits in that process.
We also want to recognise the strong work already happening in industry and the tech space, and ask an open question: is there even a need for dedicated behavioural science professionals to enter this field, or is that work already being done well by Human Computer Interaction professionals designing systems? We'll question the role automation should play in providing empathy and emotional support in care, and what that means for both patients and clinicians. We'll also be introducing AIBL, a UCL lab in health informatics working to build human-centred technology in healthcare, and hearing about the projects and collaborations they're looking for.
Panelists:
Professor Susan Shelmerdine, Associate Professor at the UCL GOS Institute for Child Health and AI advisor at The Royal College of Radiologists
Dr Juan Martin, Intensive Care Consultant at Newham University Hospital (Barts Health NHS Trust) and Associate Professor of Innovation at the University of East London
2 other speakers TBC next week!!
This event will explore questions such as:
How do clinicians build or lose trust in AI systems they didn't design and can't fully inspect, and who is responsible when those systems get it wrong?
What happens to clinical judgement, accountability, and the risk of deskilling as AI becomes part of routine decision-making?
What value can behavioural science add to health AI that isn't already being handled by IHI designers, software engineers and computer scientists?
Can and should automation provide genuine empathy and support in care, or only the appearance of it, and does that distinction matter to patients and clinicians?
What human factors and organisational barriers determine whether a validated tool actually gets used?
Are different disciplines already doing similar work under different names in developing AI-based tools for healthcare, and what would a shared language look like?
You will likely enjoy this event if you are interested in
Designing, deploying, or evaluating AI systems used in clinical settings
Understanding why technically accurate tools succeed or fail in clinical practice
Bridging behavioural science, computer science, and clinical implementation
Exploring trust, accountability, and human factors in health AI beyond the pilot stage
Event format
In-person panel discussion, hosted at UCL
Opening introduction, framing how disciplines might work together across academic institutions such as UCL and beyond, and introducing AIBL
Panel drawing on academia, computer/data science, clinical practice, and regulation, including perspectives from UCLH on hospital-level implementation
Moderated discussion followed by audience Q&A
Time for networking before and after the panel