Health Care Implications of Large Language Models like ChatGPT

Video Transcription

Video Summary

In this webinar, Linda Moy from NYU moderates a panel discussion on the use of large language models (LLMs) in radiology, highlighting efforts by the Radiological Society of North America (RSNA) to educate its members on artificial intelligence (AI) and generative AI. Panelists, including Tessa Cook, Jonathan Elias, Keith Hentel, Merrill Hussman, Felipe, and George Xie, share insights into current and potential applications of LLMs in healthcare. They discuss their experiments with internal and open-source LLMs for tasks such as clinical documentation, summarization, and patient follow-ups, while emphasizing the need for HIPAA compliance and accuracy monitoring. They also explore the role of LLMs in patient interaction, noting both potential benefits and challenges, including issues of accuracy, transparency, and regulatory approval. There is a consensus that LLMs could significantly enhance radiology practice despite current limitations and that ongoing education and transparent communication with patients are crucial. The webinar concludes with optimism for the future of AI in radiology, suggesting a promising time for new entrants in the field due to the technological advancements expected to transform practices.

Keywords

large language models

radiology

artificial intelligence

RSNA

healthcare

HIPAA compliance

patient interaction

clinical documentation

technological advancements