JAMA Perspective: AI beats doctors in cognitive tasks, human oversight not always better
HealthcareAIGuy · x · 2026-08-23
A new JAMA Perspective argues that AI already rivals or beats physicians across various cognitive medical tasks, and physician oversight may not always improve outcomes.
Key Data:
- ChatGPT o3 ranked the correct diagnosis first in 60% of complex cases vs. 15.9% for doctors.
- Another AI diagnostic system diagnosed correctly 4x more often while spending 19% less on testing.
The authors suggest that if AI is superior at a task, human supervision could degrade results, and some autonomous AI workflows could be ready by 2030. However, medicine still requires human judgment, empathy, and physical exams, and much evidence remains simulation-based.
Related event: JAMA Perspective: AI Diagnosis Outperforms Doctors(2 posts)→
More from AGI Musings
- Game AI Professor Shares Crisis of Faith on Tech's Future — Miles_Brundage · 2026-08-23
- Nobel Laureate Acemoglu: AI Future Should Be Determined by Democratic Choices — _akpiper · 2026-08-23
- Reliable 8-Hour Autonomy Matters More Economically Than the 'AGI' Label — VraserX · 2026-08-23
- AI must identify itself even if it passes the Turing test — arieljalali · 2026-08-23
- Ex-Microsoft CEO Ballmer on AGI: Humans Must Own Accountability — haider1 · 2026-08-23
- Whatnot CPO: AI Enables Senior PMs to Replace Middle Management — lennysan · 2026-08-23