JAMA+ AI Conversations

Fang Fang, professor at Karolinska Institutet leading integrative epidemiology, studies neurodegenerative disease using Nordic cohorts and AI. She discusses using AI to map disease trajectories, leverage registries and biobanks, and apply federated learning and digital twins. The conversation covers data types, ethics, generalizability, and training the next generation of epidemiologists.

John Torous, psychiatrist and digital mental health researcher, discusses AI chatbots and youth mental health. He covers safety, privacy, and legal concerns. He examines evidence standards, trial design, and benchmarking for crisis scenarios. He highlights involving patients and advocacy groups and encourages hands-on engagement to learn limitations.

Robert M. Wachter, chair of Medicine at UCSF and author of A Giant Leap, explores AI's fast arrival in healthcare. He discusses how digital progress suddenly accelerated, the rise of AI scribes and summarization tools, platform versus point-tool battles, and what clinical leaders and training programs should do next.

Adam Rodman, a general internist who leads AI programs in clinical reasoning and education, and David Wu, an MD‑PhD bridging clinical medicine and data science, discuss safe clinical use of large language models. They explore a Do No Harm benchmark and live leaderboard, failure modes like omissions, model diversity and second opinions, and how clinicians should team with AI to avoid de‑skilling.

A lively debate about studies on AI in clinical care and when machine learning can be helpful. A study on patient portal message delays raises questions about disparities and confounding. A comparison of EHR-based algorithms with in-person screening for youth suicide risk highlights limits of automated screening. A call to teach clinicians deeper critical thinking beyond pattern matching rounds out the conversation.

Frontier Large Language Model, a voice-mode LLM that narrates and analyzes AI history. It recounts ELIZA's simple therapist mimicry and PARRY's scripted paranoid persona. The discussion covers early psychotherapy scripts, ethical concerns from Weizenbaum, clashes over replacement vs augmentation, and lessons for designing humane clinical chatbots.

Dr. Cecilia Lee, a distinguished ophthalmologist from Washington University in St. Louis, delves into how retinal imaging powered by AI can unveil significant health insights. She explains oculomics and the deep connections between eye health and systemic conditions. The podcast highlights challenges like standardization and data privacy while showcasing innovative ideas such as using smartphones for widespread retinal screening. Dr. Lee also discusses the potential of retinal biomarkers in predicting cognitive decline, emphasizing the need for collaboration in this groundbreaking field.

Jonathan Chen, an Associate Professor of Medicine at Stanford, and Ethan Goh, Executive Director of the Arise Research Network, dive into the transformative landscape of AI chatbots. They discuss innovative reasoning models that enhance complex decision-making in healthcare. The duo highlights the importance of training clinicians to utilize AI effectively, while also addressing risks like automation bias. They explore why chatbots tend to agree with users and stress the need for an educational framework that prepares the next generation for AI's challenges.

Explore the rapid adoption of generative AI in U.S. hospitals and the uneven distribution across different types. Delve into the risks of early implementation and the need for robust evaluations. Learn how large vendors shape AI adoption and its impact on smaller innovators. Discover groundbreaking predictive models for mental health and the importance of validating these techniques in real-world settings. The hosts also discuss the differences in AI regulation between Europe and the U.S., and how AI can enhance clinical trial efficiency.

Eric Horvitz, Chief Scientific Officer at Microsoft and a pioneer in AI research, dives into the future of AI in medicine with Yulin Hswen. They explore how AI can safely reason in clinical settings, emphasizing the need for integrity and safety. Horvitz shares insights on the evolution of AI, from Bayesian methods to the marvels of modern generative models. The conversation also tackles public fears about AI, cognitive complacency, and the importance of calibration and reproducibility in AI systems to ensure scientific integrity. Horvitz underscores the interdisciplinary approach necessary for tackling AI's long-term societal impacts.