AI in Medicine News — 2026-11-10 (Tue)

Overview

Activity in the past 48 hours centered on diagnostics and imaging, hospital operations and clinical decision support, and drug discovery partnerships. Regulation and safety drew attention through state bans on AI therapy chatbots, a UK rethink of healthcare AI regulation, and a US approval for a heart-monitoring product. In Japan, reports focused on training, the autonomy of medical AI agents, and how evaluation data is validated.

Diagnostics and imaging

- Analysis of ECG data measured by a smartwatch may help detect reduced heart pumping function. The headline gives no accuracy figures or study stage. (www.heart.org) - A multitask AI model is reported to enhance tuberculosis diagnosis from Ugandan chest X-rays. No performance figures or evidence stage are given. (Bioengineer.org) - AI is reported to improve trainee doctors' ear disease diagnostic accuracy by 32.2 percentage points, according to AMC. The headline does not state the study design. (Korea Biomedical Review) - AI results for eight X-ray findings will reach USARAD's radiologists in their usual system. This is a workflow integration; no performance is stated. (Stock Titan) - An article on radiology AI covers its uses and rules. (CASRAI)

Clinical decision support

- Doximity expanded its clinical AI footprint to more than 190 health systems. (Fierce Healthcare) - AI helps sort specialty referrals before scheduling at UTMB. (Healthcare IT News) - AI is reported to predict PTSD risk based on stories about what brought patients to the emergency room. The headline gives no performance figures or validation stage. (Medical Xpress) - Google.org awarded Clalit $2 million to develop an AI system for patient access. (The Jerusalem Post) - Seven states ban AI therapy chatbots; the article asks what changes for clinicians. (Telehealth.org) 【Generative AI and clinical documentation】 - An article asks whether AI tools are just "EHR 2.0" and argues that history is repeating. (Medical Economics) - Dimension NXG launched an OJAS division to support hospital AI implementation. (Digital Health News) - A cancer survivor in Ramsbottom welcomed AI helping doctors at Christie focus on patients. (The Bury Times) 【Hospital operations and care delivery】 - Health system leaders discussed AI, access and a tough 2027. (Chief Healthcare Executive) - A pediatric emergency specialist is turning to AI as staffing fears deepen. (Korea Biomedical Review) - Medipal is bringing AI hospital operations automation to ASLS KOREA 2026. (헬스조선) - Barts Health will deploy AI decision support in breast cancer multidisciplinary team meetings (MDTs), described as the first live rollout. (Health Tech World) - Barnsley Tech Town AI education projects were named. (UKAuthority) 【Drug discovery and clinical trials】 No stage (preclinical, Phase I, etc.) is stated in any of the headlines below. - Quanome agreed to invest in AI drug discovery company QAI Labs. (The Quantum Insider) - Pierre Fabre and XtalPi announced an AI-driven drug discovery partnership, which includes XtalPi's AI and robotics. (Contract Pharma, BioSpectrum Asia, Unite.AI) - Inductive topped the OpenADMET CYP benchmark, its fourth straight AI drug-discovery win. This is a benchmark result, not clinical evidence. (Dealroom) 【Medical device approvals and regulation】 - A Belfast AI firm secured US approval for heart-monitoring technology. The headline does not name the product or the approving body. (Belfast News Letter) - An article covers validating AI in pharmaceutical manufacturing while maintaining compliance. It is commentary, not an approval or guidance. (Manufacturing Chemist) - The UK is rethinking healthcare AI regulation. No specific guidance is named in the headline. (Raconteur)

Safety, errors and bias

- Seven states ban AI therapy chatbots. (Telehealth.org) - AI chatbots "aren't therapists," but the article asks whether they can get men to open up. (STAT) - An article asks who decides what is "true" in medical AI. (Analytics Insight) - A ranking examines which US states are most worried about AI. (The News International) 【Research results (major journals)】 No notable developments. The only item in this category is an opinion piece titled "The Best AI Health Innovators Have a Feel for People." (American Enterprise Institute - AEI)

Developments in Japan

- Google launched an AI course of about five hours for healthcare professionals; physical therapists, occupational therapists and speech-language-hearing therapists can also take it. (PT-OT-ST.NET) - Forbes Japan reports that medical AI agents are becoming more autonomous, raising questions about safety management at medical institutions and developers. (Forbes JAPAN) - Ubie began a joint demonstration with a medical AI information platform involving Igaku-Shoin, NTT and NTT Docomo Business. (Nikkin Online) - Confirming that the evaluation data for medical AI performance is "independent" is described as important, according to Niigata University of Health and Welfare and others. (QLifePro) - GENSHI AI was commissioned to run MAGI's "medical generative AI specialist exam"; pre-orders for the official textbook have opened. (Medical Tech News)

Statements by key figures

- Geoffrey Hinton warns that an AI "Chornobyl," such as a grid brownout or a week-long financial outage, could come this year or next. The headline gives no title for him. (MLQ.ai)