Human Values in Clinical Algorithms, Old and New
Talkat Bridging Prediction and Intervention Problems in Social Systems (Simons Institute, 2026) · 2026-01-12..01-18·
summary
Lab builds an AI CPC discussant called Dr. CabOT by combining GPT-4/O3-style models, embeddings of all NEJM case records, a literature search RAG layer and text-to-speech slide generation to produce video discussants; 62 blinded comparisons judged by 10 physicians find reviewers pick the AI as the human discussant about three quarters of the time and rate its differential quality higher. Separate analysis shows recalibrating clinical equations makes 14 million Americans newly ineligible for statins. System still produces hallucinations and needs expert review and guardrails.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:14AI and clinical equations
- 3:52Recalibrating risk equations
- 10:23Clinical case conferences
- 21:11Introducing CabOT
- 28:57CabOT versus human experts
- 32:34AI progress and challenges
- 34:23Q&A
related talks
31mDo Androids Dream of Human Doctors?
21mGen AI in Health: What's New in 2026?
04mBlack Box: How AI will Amplify the Best and Worst of Humanity
28mThe Future of Healthcare – AI’s Promise and the Challenges Ahead
13mWhen AI Gets a White Coat
37mAI in medicine: where do we stand?
17mAI Innovation in Health
01h 00mKeynote Panel: AI in Healthcare