“Can ChatGPT diagnose me?” How large language models will transform clinical care

1:45 PM  |  Check-In (In person)

2:00 PM  |  Welcome & Introductory Remarks

In what ways do we see large language models impacting patient care and diagnosis? 

• Russ Altman - Kenneth Fong Professor and Professor of Bioengineering, Stanford University
• Curtis Langlotz - Professor of Radiology, Director of the Stanford Center for Artificial Intelligence in Medicine and Imaging, Stanford University

2:15 PM  |  Large Language Models 101

What are LLMs? How do they work? What data were they trained on? How should we think about general purpose LLMs like ChatGPT vs health-specific Foundation Models? 
What are some of the key limitations in its training and development that consumers and clinicians should be aware of when using it for medical advice? 

• Nigam Shah - Professor of Medicine and of Biomedical Data Science, Co-Director of the Stanford Center for Artificial Intelligence in Medicine and Imaging, Stanford University

2:30 PM  |  Panel 1: Direct-to-Consumer Uses of ChatGPT/LLMs: How might LLMs transform a consumer’s ability to better diagnose and manage their health conditions?

A high percentage of patients currently use online search engines to self-diagnose when they experience new symptoms. In what ways does ChatGPT/LLMs change how a patient interacts with online health information to make decisions about whether to seek care or not? 

• Daniel Yang - Program Director, Gordon and Betty Moore Foundation (Moderator)
• Andrea Downing - Founder, The Light Collective
• Eric Horvitz - Chief Scientific Officer, Microsoft
• Nabiha Syed - Chief Executive Officer, The Markup

3:00 PM  |  Panel 2: How might LLMs/ChatGPT support clinicians in diagnostic decision-making? 

In what ways might these LLMs support or even replace clinicians in the diagnostic process? What types of products would clinicians find useful? What are some of the pitfalls that clinicians need to be wary of when using these tools? 

• Curtis Langlotz - Professor of Radiology, Director of the Stanford Center for Artificial Intelligence in Medicine and Imaging, Stanford University
• Matthew Lungren - Chief Medical Information Officer, Nuance
• Nigam Shah - Professor of Medicine and of Biomedical Data Science, Co-Director of the Stanford Center for Artificial Intelligence in Medicine and Imaging, Stanford University
• Ida Sim - Professor of Medicine and Computational Precision Health, Chief Research Informatics Officer, UCSF

3:30 PM  |  Panel 3: The Ethics, Equity, and Regulation around LLMs/ChatGPT

In what ways might these new tools affect diagnostic disparities in the US? How might the pre-existing biases in the training data potentially perpetuate disparities? How might we mitigate these effects? How should these tools be classified from a regulatory perspective (e.g., as a medical device or as a general wellness tool)? And how might that influence their safety, efficacy and usefulness? Other ethical considerations (e.g., patient consent, data privacy, risk of security / adversarial attacks)

• Justin Norden - Partner, GSR Ventures (Moderator)
• Glenn Cohen - Deputy Dean and James A. Attwood and Leslie Williams Professor of Law, Harvard Law School; Faculty Director, Petrie-Flom Center for Health Law Policy, Biotechnology & Bioethics, Harvard University
• Lisa Lehmann - Brigham and Women's Hospital, Department of Medicine and Harvard Medical School
• Yauheni Solad - Chief Medical Information Officer, UC Davis Health
• Troy Tazbaz - Director of Digital Health Center of Excellence, FDA

4:00 PM  |  Networking Reception (In person)