JAMA +AI Weekly Update January 3, 2026

AI chatbots have advanced rapidly, incorporating new reasoning architectures that reshape decision-making and medical education.

Jonathan Chen, MD, PhD, and Ethan Goh, MD, MS, of Stanford University join JAMA and JAMA+ AI Associate Editor Yulin Hswen, ScD, MPH, to discuss the latest generation of AI models, the importance of evaluating benefits and harms, and sycophancy in AI systems.

Learn more and listen to the podcast.

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

• A multicenter study of patients with advanced melanoma found that higher levels of tumor-infiltrating lymphocytes (TILs), quantified by a publicly available AI model on routine histology slides, were independently associated with better response and survival after immune checkpoint inhibitor therapy. These findings suggest AI-driven TIL scoring could serve as an accessible biomarker for immune checkpoint inhibition outcomes. (JAMA Oncology)
• An AI-driven pipeline using pretreatment CT scans accurately detected imaging-based extranodal extension (iENE), a marker linked to worse outcomes in HPV-associated oropharyngeal carcinoma. AI-predicted iENE outperformed expert radiologists and was independently associated with lower overall and recurrence-free survival, suggesting a potential tool to complement clinical decision-making and support institutions without specialized imaging expertise. (JAMA Otolaryngology–Head & Neck Surgery)
• A cross-sectional study of 514 hospitals across 17 states found that while nearly all offered patient portals, only 10.9% provided access in three or more languages, and only 4.7% included the most common non-English, non-Spanish language in their county. Teaching hospitals were more likely to offer multilingual access, while portals from major vendors like Epic were associated with better accessibility. (JAMA Network Open)