The Africa Health-Tech Accelerator's 2026 cohort targets AI diagnostics and supply chains alike, reflecting growing but still modest investment in African health tech relative to funding rounds elsewhere.
A new analysis finds most health plans now use AI for prior authorization approvals, but a meaningful share also use it for denials — a distinction advocates say current federal AI policy dangerously fails to make.
Stanford's new Virtual Biotech platform aims to compress drug discovery timelines through AI simulation, an academic counterpart to the wave of commercial AI drug discovery deals reshaping pharma research.
AstraZeneca's up to $1.46 billion deal with insitro bets on an AI platform that grows lab cells to uncover the causal biology behind fatty liver disease, though the partnership remains at its earliest, highest-risk stage.
Two large Chinese clinical trials — one on rare disease diagnosis, one an 8,240-patient orthopedic study — offer some of the most rigorous real-world evidence yet on AI's clinical impact, with results still pending.
Wellysis's FDA-cleared S-Patch CardioAI joins a list of more than 1,600 AI-enabled medical devices, even as the FDA weighs new rules specifically for generative AI devices that could reshape the approval pathway.
Confido Health's $12.5 million extension targets AI for medical practice operations, a less flashy but steadily funded corner of health AI focused on scheduling, billing and administrative efficiency.
Teladoc's new hospital ambient scribe and AKASA's autonomous coding platform both push AI documentation tools into higher-stakes settings with less human oversight than earlier assistive versions.
The UK's Medical Protection Society says AI-generated patient complaints, sometimes exceeding 20 pages, are adding to GPs' administrative burden even as AI tools simultaneously ease their documentation workload elsewhere.
South Korea's new guidelines target emotional dependence on AI chatbots and risky self-diagnosis, as global regulators grapple with mental health AI tools that research shows can both help and harm vulnerable users.