Wrong AI advice hurts doctors' judgment more than right advice helps: study of 16,641 decisions
ValerioCapraro · x · 2026-11-16
A new analysis of 16,641 clinical decisions finds an asymmetry in how AI support affects doctors interpreting MRI scans, ECGs, X-rays and endoscopy videos. Doctors first judged each case independently, then saw AI advice and could revise.
- When AI was correct, doctors' accuracy rose by 5.8 percentage points.
- When AI was wrong, accuracy fell by 7.6 points — a third larger than the gain.
- Overall accuracy still improved from 76.5% to 79.7%, since AI was right about 81% of the time.
Worryingly, doctors sometimes abandoned correct judgments after seeing incorrect AI advice. The author argues we need methods to help users distinguish advice worth following from advice worth rejecting — knowing when to reject AI advice is part of using it well.
More from AGI Musings
- US CS enrollment down 18.4% in two years as one bet contrarians will win the AI era — bendee983 · 2026-11-16
- Rutger Bregman: AI denial is the left's new climate denial — and more dangerous — marigo · 2026-11-16
- Terence Tao: math has moved from proof scarcity to proof abundance — haider1 · 2026-11-16
- 'We taught AI to write, paint and compose — the response has been surprisingly medieval' — TinfoilTricorn · 2026-11-16
- Economist Daniel Gross: defining AI-era 'scientific flourishing' is an epistemological minefield — soumitrashukla9 · 2026-11-16
- Dev: more worried about AI builders justifying harm 'for the greater good' than AI doom — Dan_Jeffries1 · 2026-11-16