

AI, health care, and the strange future of medicine
May 28, 2024
Physician panel discusses AI in healthcare, covering consumer symptom assessment apps, generative AI tools, population health data science, and AI integration in emergency medicine. They explore challenges like bias concerns, data regulations, and patient empowerment with AI tools. The discussion emphasizes human oversight in medical decisions and the critical need for transparency and consent in healthcare data usage.
AI Snips
Chapters
Transcript
Episode notes
Evolution and Challenges of AI in Healthcare
- AI tools in healthcare have evolved from simple point solutions to more complex, general-purpose tools like generative AI.
- While promising, generative AI presents challenges like hallucinations and non-deterministic outputs, necessitating a focus on safety and clinician partnership.
Hidden Risk in EHR
- Jesse Ehrenfeld recounts a patient's near-miss cardiac arrest during a prior surgery.
- This critical information was buried in a nursing note, highlighting EHR usability issues.
Patient-Facing AI
- Claire Novorol shares how Ada Health pivoted to a patient-facing model.
- Initially designed for doctors, time constraints led them to gather patient information upfront, improving efficiency and patient engagement.