AI doesn’t just get things wrong. It gets things wrong with complete confidence — citing studies that don’t exist, building logical arguments around biological impossibilities, and delivering dangerous recommendations in the fluent, authoritative voice of a clinical expert. In healthcare, that gap between confidence and accuracy isn’t an inconvenience. It’s a patient safety crisis.

In Episode 3 of AI Literacy for Leaders, Laurence Gill breaks down the mechanics of AI hallucination in clinical settings — drawing on guidance from the World Health Organization, UK regulators, and a landmark study that tested whether humans can actually catch AI lies. The findings are more alarming than most healthcare leaders realize.

You’ll learn the three distinct types of AI hallucination that clinicians need to recognize, why experienced physicians miss them more often than you’d expect, and why using a second AI to check the first one doesn’t solve the problem. You’ll get a practical green-yellow-red framework for where AI is safe to use, where it requires careful oversight, and where it should never go near a clinical decision. And you’ll hear about a failure mode that almost nobody is talking about — not AI that lies, but AI that goes dangerously silent.
The episode uses storylines from The Pitt on HBO Max — where a hospital’s AI clinical assistant hallucinates a treatment recommendation and calls the entire tool into question — as a narrative anchor for what real healthcare organizations are navigating right now.

The future of medicine isn’t AI versus doctor. It’s the clinician who knows how to interrogate AI output versus the one who accepts it at face value. This episode gives you the framework to be the former.- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Runtime: Approx. 30 minutes

Hosted by: Laurence Gill

Series: AI Literacy for Leaders