The Promise and Pitfalls of AI in Medicine: Bob Wachter
Apr 18, 2024
auto_awesome
Bob Wachter, Chair of Medicine at UCSF, discusses the promise and pitfalls of AI in healthcare. Topics include AI's empathy, predictive models, consolidation of EHRs, de-skilling of clinicians, use cases like digital scribes, concerns about bias, and the role of AI in medicine. The conversation explores trust, accuracy, and human oversight in AI-generated medical content.
AI can outperform human clinicians in empathy, transforming patient interactions.
Predictive models in healthcare show versatility, but translating minor accuracy gains to clinical significance is a challenge.
Deep dives
AI and Medicine: Current Applications and Concerns
AI in medicine poses a dichotomy between short and long-term impacts. Short-term benefits include enhancing healthcare quality, access, and efficiency, especially in addressing prevailing healthcare system flaws. Despite concerns about job displacement and misinformation risks, AI shows promise in improving patient outcomes and clinician practices.
Predictive Models: Evolution and Applications
Initial skepticism of predictive models has shifted towards recognition of their versatility in various healthcare scenarios. While accuracy improvements are noted, translating minor accuracy upgrades into clinical significance raises questions. Applications extend beyond traditional predictions to real-time decision support, augmenting clinical decision-making.
Empathy in AI Interaction and Clinical Practice
AI prowess in displaying empathy challenges traditional assumptions about human-to-AI interactions. Studies revealing AI's superior empathy to clinicians in patient interactions evoke reevaluation of AI's communication potential. As AI evolves, its capacity to enhance clinical compassion and communication skills remains a focal point.
Operational and Clinical AI Integration: Advancements and Future Scenarios
Digital scribes emerge as front-runners in AI integration, streamlining documentation processes and enabling efficient care practices. Anticipated applications range from chart summarization to identifying hospital-at-home eligibility criteria. Striking a balance between de-skilling inefficiencies and enhanced clinical performance remains pivotal for AI integration success.
Eric asks the question that is on many of our minds - is the future of AI more Skynet from Terminator, in which AI takes over the world and drives humanity to the brink of extinction, or Wall-E, in which a benevolent and empathetic AI restores our humanity?
Findings that in several studies AI was rated by patients as more empathetic than human clinicians (not less, that isn’t a typo). Turns my concern about lack of empathy from AI on its head - the AI may be more empathetic than clinicians, not less.
Skepticism on the ability of predictive models to transform healthcare
Consolidation of EHR’s into the hands of a very few companies, and potential for the drug and device industry to influence care delivery by tweaking AI in ways that are not transparent and already a sort of magical black box.
AI may de-skill clinicians in the same way that autopilot deskilled pilots, who no longer new how to fly the plane without autopilot
A live demonstration of AI breaking a cancer diagnosis to a young adult with kids (VITAL Talk watch out)
Use cases in healthcare: Bob predicts everyone will use digital scribes to chart within two years
Concerns about bias and other anticipated and unanticipated issues
And a real treat- Bob plays the song for this one! Terrific rendition of Tomorrow from the musical Annie on piano (a strong hint there about Bob’s answer to Eric’s first question). Enjoy!
-@AlexSmithMD
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode