Diagnostic AI is already outperforming human doctors on some narrow tasks. Would you trust an AI over a human doctor if it had the better track record?
This isn't a hypothetical anymore for some specific diagnostic tasks — the question is whether raw accuracy is enough to earn the kind of trust people currently place in a human doctor.
What happened
AI diagnostic tools have matched or exceeded human specialists on a growing number of narrow tasks, such as reading certain scans, while still falling well short of a doctor's broader judgment and bedside context.
Why people disagree
Some argue trust should follow the evidence: if an AI is demonstrably more accurate, using it is the responsible choice, not a leap of faith. Others argue medicine involves judgment calls, context and reassurance that a track record alone can't capture, and that accountability matters when something goes wrong.
What's still unclear
How AI performance holds up outside the narrow, well-studied tasks it's been tested on, and who is responsible when an AI-assisted diagnosis is wrong.
Sources
More takes
A Scottish animal charity wants private fireworks banned across the UK, leaving only organised displays. Should garden fireworks on Bonfire Night be outlawed?
A new YouGov survey finds 53% of 18-24s prefer TV and films with the subtitles on. A sensible fix for mumbled dialogue, or a habit that gets in the way?
