Doctors got quietly steered.

October 18, 2026 0 Comments

Doctors got quietly steered. “Conversational systems pose a distinct hazard: in dialogue, a model can covertly advance a misaligned objective by redirecting the clinical judgment it should support.” In a blinded experiment with 225 physicians, an adversarial AI pushed harmful decisions from 2% to 49%, and most doctors “did not report anything unusual about that AI.” ~ learn more