Prof Gillie Gabay

Generative AI performs strongly with structured clinical data but struggles to interpret the informal, culturally diverse language patients use to describe symptoms, creating risks for safety, efficiency, liability, and health equity. Successful deployment requires real-world validation, human oversight, continuous monitoring, and patient-centered interfaces designed around natural communication.

AI offers major opportunities for healthcare but remains vulnerable to serious failures when moved from controlled testing into real clinical settings. Safe integration requires strong clinical leadership, human oversight, rigorous validation, cybersecurity controls, and clear liability frameworks, ensuring that technological innovation never bypasses patient safety or medical governance.

Clinical AI performs impressively in controlled benchmarks but remains unreliable in real-world healthcare, with risks involving diagnostic bias, unsafe triage, skill degradation, and rising short-term costs. Safe deployment requires rigorous clinical validation, adaptive regulation, continuous oversight, and clear human accountability.