JAMA + AI Weekly Update FEBRUARY 14, 2026

How do we ensure AI actually helps clinicians deliver better care? In the latest episode of JAMA+ AI Conversations, Editor in Chief Dr Perlis, MD, MSc, and Associate Editor Yulin Hswen. ScD, MPH, examine this question through the lens of three recent studies drawn from the JAMA Network.

Roy shares why each article makes him nervous about the potential for AI to adversely impact patient care. Could AI help reduce disparities in responses to patient messages, or will it exacerbate these differences? Are AI tools better at screening for suicide risk, or will they lull us into a sense of false confidence? Additionally, why are particular kinds of critical thinking skills more important than ever?

These studies press us to scrutinize the promises and perils of AI-driven care.

Learn more and listen to the podcast.

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

• A cohort study found that a suicide risk algorithm built from electronic health records identified pediatric emergency department patients at risk of subsequent suicide attempts more accurately than universal in-person screening tools. The analysis included nearly 20,000 youths aged 10 to 18.
• A Viewpoint examines how strengthening critical thinking grounded in human pathophysiology could improve diagnostic reasoning, in contrast with traditional approaches centered on illness scripts. Authors argue that moving beyond pattern recognition is essential as AI shapes clinical evaluation.
• A cross-sectional analysis of more than 1.2 million primary care message threads found slower response times for Black, Hispanic, Medicaid/Medicare dual‑eligible, and Spanish‑preferring patients, compared with White, commercially insured, and English‑preferring patients.