Healthcare AI trust crisis
U.S. medicine is facing a mounting AI trust crisis — commentators say adoption must “move at the speed of trust, not investment,” warning patients and clinicians about opaque algorithmic decisions (statnews.com). European leaders push the same alarm: a HealthAI CEO told Euronews AI could avert system collapse but only with careful rollout and oversight (euronews.com).
A 50‑state survey that pooled 582,634 responses from 443,455 unique U.S. adults found the share reporting “a lot of trust” in physicians and hospitals fell from 71.5% in April 2020 to 40.1% in January 2024. (JAMA Network Open — ) An independent safety evaluation published in Nature Medicine tested ChatGPT Health on 60 realistic clinical scenarios and found it under‑triaged roughly half of cases that clinicians judged to be emergencies, advising delayed care in a majority of those instances. (Nature Medicine — ) The American Medical Association’s March 12, 2026 survey reported AI use among physicians roughly doubled since 2023, with 81% of respondents saying they use AI in practice, even as other surveys show clinicians express persistent accuracy and reliability concerns. (American Medical Association — ) Doximity’s 2026 State of AI in Medicine report found 94% of physicians are using or interested in AI but 71% worry about accuracy and reliability, highlighting a split between rapid clinical adoption and lingering professional skepticism. (Doximity / BusinessWire — ) About 2,400 mental‑health clinicians at Kaiser Permanente staged a one‑day strike in Northern California in mid‑March 2026, citing fears that AI‑driven screening and triage could replace human assessment and miss patients at high risk of self‑harm. (Associated Press — ) Dr. Ricardo Baptista Leite, CEO of HealthAI, told Euronews on March 24, 2026 that AI presents an opportunity to expand access to care but that systems must be reconfigured and tightly governed for it to prevent health‑system collapse. (Euronews — ) European regulatory frameworks are tightening: the European Health Data Space entered into force in 2025 to standardize data reuse across the EU, and the EU AI Act’s high‑risk requirements — including Article 12 logging and traceability obligations — become fully applicable to medical AI in 2026, raising compliance milestones for vendors and health systems. (European Commission — ) (GLACIS — )
