Yale Certificate in Medical Software and Medical AI: Guest Experts

Large Language Model-based Chatbots and Medical Regulation

Apr 16, 2024

Prof. Stephen Gilbert, a leading expert in Medical Device Regulatory Science, dives into the intersection of AI and medical regulation. He discusses the challenges of adapting traditional frameworks for AI chatbots and large language models, emphasizing the need for updated regulations to ensure patient safety. Gilbert compares LLMs to an unpredictable genie, highlighting their potential and risks. He also explores the evolving responsibilities of physicians in a tech-driven landscape, advocating for collaboration to advance innovations while maintaining safety.

Ask episode

Chapters

Transcript

Episode notes

Intro

00:00 • 5min

Navigating Regulation in Medical AI Chatbots

04:38 • 3min

Regulating the Genie: Large Language Models in Medicine

07:44 • 27min

Exploring the Intelligence of Medical Chatbots

35:10 • 3min

Navigating AI Regulatory Challenges

37:48 • 12min

Navigating AI in Healthcare: The Physician's Dilemma

49:39 • 10min

Our guest is Prof. Stephen Gilbert (https://www.linkedin.com/in/stephen-gilbert-31ba2587/) who is a Professor of Medical Device Regulatory Science at the Else Kröner Fresenius Center for Digital Health, Technische Universität Dresden where he teaches and conducts research on regulatory science with a team of colleagues. He is also News and Views Editor, Nature Portfolio – Digital Health. He worked in senior MedTech and Digital Heath roles in industry for 5 years, before returning to academia in 2022.

His research goals are to advance the regulatory science of software as a medical device and AI-enabled medical devices. Innovative digital approaches to healthcare must be accompanied by innovative approaches in regulation to ensure speed to market, to maximum access of patients to life saving treatments whilst ensuring safety on market. His main research interests are in: (i) data sharing and the European Health Data Space; (ii) approaches to market approval of adaptive AI enabled medical devices; (iii) drugdigital/AI-enabled medical device product realisation; (iv) digital/virtual twins: as an organising concept of the future of healthcare.”

Further Reading

Derraz B, Breda G, Kaempf C, Baenke F, Cotte F, Reiche K, Köhl U, Kather JN, Eskenazy D, Gilbert S. New regulatory thinking is needed for AI-based personalised drug and cell therapies in precision oncology. NPJ Precis Oncol [Internet]. Nature Publishing Group; 2024 Jan 30 [cited 2024 Jan 30];8(1):1–11. Available from: https://www.nature.com/articles/s41698-024-00517-w
Gilbert S, Harvey H, Melvin T, Vollebregt E, Wicks P. Large language model AI chatbots require approval as medical devices. Nat Med [Internet]. Nature Publishing Group; 2023 Jun 30 [cited 2023 Jun 30];1–3. Available from: https://www.nature.com/articles/s41591-023-02412-6

Gilbert S and Kather JN. Guardrails for the use of generalist AI in cancer care. Nature Reviews Cancer [Internet]. Nature Publishing Group; 2024 Apr 16 [cited 2024 Apr 16]. Available from: https://www.nature.com/articles/s41568-024-00685-8