Can Open-Source LLMs Compete With Proprietary Ones for Complex Diagnoses?

In this follow-up to a 2017 interview with JAMA Medical News, the University of Southern California’s Maja Matarić, PhD, the computer scientist who pioneered the field of socially assistive robotics, discusses how artificial intelligence is advancing the field in areas ranging from autism to physical rehabilitation to anxiety and depression. Related Content: Social Robots That Help Support People’s Health Are Getting a Boost From AI Socially Assistive Robots

Kimmie Ng, MD, MPH, a JAMA associate editor and associate chief of the Division of Gastrointestinal Oncology at the Dana-Farber Cancer Institute, shares highlights from the American Society of Clinical Oncology’s annual meeting, including new research on diet, exercise, and cancer survival and the best time of day for treatment. Related Content: Lifestyle and Cancer Survival, the Best Time of Day for Treatment, and More—Highlights From ASCO

New NIH-FDA Partnership Targets Nutrition Research Gaps; First Blood Test for Alzheimer Biomarkers Receives FDA Clearance; A Growing Movement to Care for Caregivers Related Content: New Federal Program Seeks to Bridge Nutrition Research With Regulatory Policy What to Know About the First FDA-Cleared Blood Test for Alzheimer Biomarkers As the US Ages, a Growing Movement Aims to Care for Caregivers

Delaying diagnosis of parkinsonism can mean delaying care. In a study recently published in JAMA Neurology, David Vaillancourt, PhD, and colleagues tested the ability of an AI model to differentiate between Parkinson disease and other neurodegenerative disorders when paired with MRI. He joins JAMA and JAMA+ AI Associate Editor Yulin Hswen, ScD, MPH to discuss. Related Content: A Large Proportion of Parkinson Disease Diagnoses Are Wrong—Here’s How AI Could Help Automated Imaging Differentiation for Parkinsonism

Federal Funds for Rural Health Care May Be Cut; Why the IV Fluid Shortage After Hurricane Helene Was Years in the Making; Surge in US Sports Betting Raises Public Health Concerns Related Content: Federal Funds for Rural Health Care Are on the Chopping Block—Here’s What That Could Mean IV Fluid Shortages Persist Months After Hurricane Helene Hit a Supplier—Hospitals Have Had to Adapt The Hidden Health Costs Associated With Legalized Sports Gambling

Susan Athey, PhD, of Standford University joins JAMA+ AI Editor in Chief Roy H. Perlis, MD, MSc, to discuss her research on machine learning to target behavioral nudges for college students and their potential implications for health care. Related Content: How an Economist’s Application of Machine Learning to Target Nudges Applies to Precision Medicine

Ajay J. Kirtane, MD, SM, a professor of medicine at Columbia University Irving Medical Center, discusses late-breaking clinical research presented at the annual meeting of the American College of Cardiology in an interview with JAMA Medical News Director Jennifer Abbasi. Related Content: Heart Health Highlights From ACC—Marathon Runners and Mortality, Oral GLP-1 Drugs, and More

Axe Falls on Federal Health Workforce; Experts Say CDC Cuts Will Cost Lives Related Content: “Guaranteed Pandemonium” as HHS Secretary Slashes Federal Health Workforce Experts Say Abrupt and “Staggering” CDC Cuts Will Cost Lives

Diabetic retinopathy remains a leading cause of preventable blindness worldwide, and AI may facilitate screening, if such models continue to perform well when they are deployed in the real world. Coauthors Arthur Brant, MD, of Stanford University, and Sunny Virmani, MS, of Google join JAMA+ AI Editor in Chief Roy H. Perlis, MD, MSc, to discuss a new study published in JAMA Network Open. Related Content: Diabetic Retinopathy Is Massively Underscreened—an AI System Could Help Performance of a Deep Learning Diabetic Retinopathy Algorithm in India

A recent study published in JAMA Health Forum suggests that institutions may be able to deploy custom open-source large language models (LLMs) that run locally without sacrificing data privacy or flexibility. Coauthors Thomas A. Buckley, BS, and Arjun K. Manrai, PhD, from the Department of Biomedical Informatics at Harvard Medical School join JAMA+ AI Editor in Chief Roy H. Perlis, MD, MSc, to discuss. Related Content: Can Open-Source AI Models Diagnose Complex Cases as Well as GPT-4?