OpenAI’s Head of Health AI on Whether You Should Trust ChatGPT

a16z

Summary

This discussion explores how OpenAI is integrating health intelligence into its general-purpose models, focusing on balancing safety, accuracy, and accessibility. By leveraging a global network of medical experts to define behavior and evaluation metrics, OpenAI aims to provide a reliable 'health co-pilot' for patients and clinicians alike. The conversation covers the transition from health information to integrated care, the importance of grounding responses in medical literature, and a future-looking perspective where AI reduces the administrative friction of the healthcare system, ultimately aiming for more human-centric clinical experiences.

Chapters

  1. 0:04Addressing scarcity in healthcare
  2. 0:07Mission and the Consumer Health Landscape
  3. 0:21Real-world patient impact
  4. 1:44Why general models for health
  5. 2:29Usage trends in ChatGPT
  6. 6:38Architectural Control and Model Training
  7. 7:23Avoiding the obsolescence of smaller models
  8. 7:40The benefit of scale
  9. 8:48The Future of Proactive Care
  10. 11:19Clinician co-pilots
  11. 12:13The role of AI agents
  12. 16:12Trust, Disagreement, and Safety
  13. 17:30Navigating uncertain cases
  14. 18:13Prioritizing safety over consumer comfort
  15. 22:32Managing medical uncertainty
  16. 26:23Integrating AI into Clinical Workflows
  17. 26:39Collaboration with EHR systems
  18. 34:17Reducing administrative burden
  19. 36:45The personal health assistant
  20. 41:23AI in medical research

Topics covered

  • Mission and Vision
  • Consumer Health Usage

+5 more topics (full content when you follow the channel)

Transcript excerpt

5.3 plus trillion we spend on health care. If it were its own country, it'd be the third largest GDP. And yet, we have a ton of scarcity. >> OpenAI's mission is to ensure AGI benefits all of humanity and give people access to trustworthy, abundant, reliable health intelligence. One of the stories that I was most touched by was the story of Nathan Labenz. He basically had this kind of long saga with his son where at some point he was suffering a lot of fatigue and irritability and AI helped them navigate the process and it turns out he was diagnosed with an aggressive form of leukemia. >> Some of the work that you've done around mental health, there's a lot of complexity. How do you help guide a user on that given that could mean so many things? >> We've just been working with a bunch of licensed mental health experts across the world that think about the right ways to define model behavi…