Medical AI reasoning collapses under scrutiny
A new benchmark reveals that top medical AI models look accurate on the surface but fail at consistent logical reasoning nearly half the time.
A new benchmark reveals that top medical AI models look accurate on the surface but fail at consistent logical reasoning nearly half the time.