Physician-Style Licensure for Clinical AI? A 14-Post Thread Argues Trust Can Be Manufactured
jonc101x · x · 2026-09-09
Physician Jonathan posted a 14-part thread arguing that letting autonomous AI systems prescribe medications without human double-checking—once too scary to say aloud—now seems obvious and inevitable.
Key points:
- LLMs have already crushed written multiple-choice exams: "This game is over," but goalposts keep moving because unlimited knowledge backed by zero responsibility does not earn trust.
- Borrowing from human physician licensing, he proposes a "Medical (AI) Licensure" pipeline as trust-manufacturing: certify and hold autonomous systems accountable the way doctors are.
- This isn't hypothetical: Utah's regulatory sandbox already allows companies like Doctronic to do autonomous medication refills.
- The thread closes by pointing to his BMJ Digital Health & AI piece with Michael Pfeffer and CALonghurst on why humans remain in the loop, plus an FDA discussion paper requesting public feedback.
Bottom line: the bottleneck for clinical AI is not capability but institutionalized trust, and sandbox programs plus licensure-style certification may be the path forward.
More from AGI Musings
- Stanford-Harvard ARISE releases inaugural State of Clinical AI Report 2026 — jonc101x · 2026-09-09
- Anthropic researcher: without AI, US GDP growth would be ~1% — QuintinPope5 · 2026-09-09
- Ben Todd: if alignment experts say it could kill everyone, react with "oh shit", not PR critiques — sjgadler · 2026-09-09
- Is 'Pointing Enough ChatGPT Agents at Navier-Stokes' an Extreme Idea? — teortaxesTex · 2026-09-09
- "Change Could Come Any Day": Poster Distrusts OpenAI and Anthropic on Alignment — justalexoki · 2026-09-09
- tszzl: solving mechinterp within a year could halve his p(doom); utopia is seriously possible — tszzl · 2026-09-09