JAMA + AI Weekly Update November 8, 2025

How is AI redefining the process of scientific discovery for biotech and pharma? In this week’s JAMA+ AI Conversations, JAMA+ AI Editor in Chief Roy Perlis, MD, MSc, interviews Ajamete Kaykas, PhD, Chief eXploration Officer at insitro, a biotech company aiming to combine machine learning with wet lab research to accelerate new therapeutic development.

They discuss the future of scientists in the AI era and whether AI will open the door to a new generation of therapeutics. Don’t miss this insightful conversation at the intersection of technology, data science, and medicine.

Learn more and listen to the podcast.

Editor’s Picks in this week’s JAMA+ AI:   

• For very preterm infants with severe thrombocytopenia, current evidence shows that prophylactic platelet transfusions do not uniformly reduce the risk of major bleeding or mortality. This modeling study found substantial variation among individuals in predicted benefits and harms of prophylactic platelet transfusion based on their current clinical characteristics. (JAMA)
• In an analysis using FDA data, manufacturers of more than 90% of FDA-listed AI- and machine learning–enabled devices publicly disclosed no payments to clinicians or teaching hospitals from 2017-2023. In the same time frame, the dollar amount of general payments doubled, reflecting gaps in reporting and regulatory classification. (JAMA)
• Researchers compared the accuracy of AI-based translation models with professional human translators for pediatric inpatient discharge instructions in Spanish, Simplified Chinese, Somali, and Vietnamese. AI translation was comparable to professional translation for Spanish, but performed worse for Chinese, Vietnamese, and Somali. (JAMA Network Open)