Large multimodal models jump from roughly 60% to 80% accuracy on neuroradiology cases when answer choices are provided.
US patent law still demands a human inventor even when AI did the creative work.
The FDA's new discussion paper proposes doctor-style competency testing for generative AI medical devices before they reach patients.
Hippocratic AI launches Agentic Orchestrators that supervise teams of voice AI agents aimed at measurable healthcare outcomes.
OpenAI's nonprofit arm commits $100 million to help states apply AI tools to hepatitis C care and other public health…
Oracle Health's OpenAI-powered patient portal is generally available in the U.S., translating medical jargon into plain language.
MIT Technology Review explores how AstraZeneca and others use AI to compress drug discovery and target previously untreatable diseases.
The Coalition for Health AI launches a national initiative with enterprise AI licenses donated by OpenAI and Anthropic.