Chasing Life AI Is in Your Healthcare Now. Here’s What to Know
27 snips
May 8, 2026 
Dr. Bob Wachter, a physician and author who studies digital health, explores how AI is already used in care and where it might go wrong. He discusses AI scribes, clinician tools like Open Evidence, patient-facing apps, privacy tradeoffs, and which tasks should remain human-led. The conversation balances optimism for improved access with caution about safety, trust, and training.
AI Snips
Chapters
Books
Transcript
Episode notes
Machines Get Harsher Judgment Than Humans
- People hold higher expectations and lower trust for machine errors compared with human errors, affecting acceptance.
- Wachter compares Waymo's safety data to human drivers to show machine trust can be merited despite higher scrutiny.
Use AI To Automate Routine Clinical Tasks
- Delegate routine tasks to AI where safe, like medication refills, to free clinicians for complex care.
- Wachter highlights Utah allowing an AI to autonomously refill ~400 common meds as a first-step experiment.
AI Performance Depends On User Expertise
- The same AI performs very differently for experts versus lay users because experts know better prompts and can interpret nuance.
- Wachter says tools that work for doctors may fail for patients who lack medical literacy and need more guided questioning.
