STAT AI: An Emergency Medicine Podcast · Ethan Abbott and DJ Apakama

EP7 Nick Genes, MD, PhD, Emergency Medicine, NYU Langone (2nd half of conversation)

·30 min·2 clips

Dr. Nick Genes reveals why AI in healthcare is now entering the trough of disillusionment after last year's hype peak.

This episode continues a conversation with Dr. Nick Genes, an emergency physician and informaticist from NYU Langone. Hosts DJ Abhakama and Ethan Abbott, both emergency physicians, explore the integration of artificial intelligence into their specialty. Dr. Genes provides a mid-career perspective on technological shifts in medicine, comparing the current AI momentum to past changes like EHR adoption. He describes the AI landscape as having reached a "peak of inflated expectations" in 2023, now entering a "trough of disillusionment" as implementation challenges and bias issues become apparent. Dr. Genes recalls early resistance to electronic health records, where some clinicians felt order sets and computerized decision support undermined their professional identity. He argues that AI's current strengths, like summarizing patient data for handoffs, similarly touch on tasks doctors view as core to their role. The discussion covers medical education, where Dr. Genes advocates for students to learn foundational clinical reasoning without AI reliance, similar to delaying screen time for children. He emphasizes that informatics training involves change management and consensus-building more than programming. Dr. Genes expresses concern that generative AI's rise, led by large tech companies, complicates ethical oversight and bias detection compared to earlier, more transparent machine learning models. He cites Google DeepMind research highlighting subtle biases in large language models that require extensive, ongoing evaluation. A practical challenge raised is whether clinicians should disclose to patients which AI-generated alerts they acted upon or ignored, such as for preventative colonoscopy screenings. Dr. Genes is personally excited by AI applications that optimize emergency department operations, like prioritizing radiology studies for fast-track patients. He references the ambitious NYU Tron project, which uses a large language model as an all-purpose prediction engine for hospital-wide decision-making. For clinicians starting with AI, he recommends viewing it as a collaborative partnership rather than a transactional tool and using techniques like "chain of thought" prompting to understand its reasoning. Dr. Genes also discusses his role on an ACEP AI task force, which plans to survey emergency physicians about their attitudes and concerns regarding AI's impact on their work and training. The tone is conversational and educational, featuring a knowledgeable guest sharing insights and historical context with two engaged hosts. The style is exploratory, balancing excitement about AI's potential with pragmatic caution about its ethical and practical pitfalls. This episode would appeal to emergency medicine clinicians, medical educators, and informatics professionals interested in the realistic near-term future of AI in healthcare. Listeners seeking highly technical AI details or definitive predictions about long-term disruption might find the discussion too cautious or broad.

As heard by us

A grounded look at how emergency medicine AI is becoming more opaque as generative systems move in.

Dr. Nick Jeans' conversation with Stat AI frames a clear shift in emergency medicine AI: the earlier wave of smaller, more legible tools could be tested in a defined population and explained to clinicians, while generative AI is bringing in larger systems that are harder to see…

Read the full review in PlayNext →

Why you'd press play

When you want a measured take on AI in emergency medicine, with bias and reproducibility kept front and center.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from STAT AI: An Emergency Medicine Podcast