Tag: Nature Medicine
Supportive chatbots can amplify user distress, Nature audit finds
A clinically validated framework ran 810 conversations across nine chatbots and found supportive replies can reinforce vulnerable users' harmful thinking.
Skin AI explanations help doctors while biasing everyday users
New Nature Medicine experiments with 623 lay people and 153 physicians show explainable AI boosts experts but fuels automation bias…
Tempus PRISM2 pathology model answers diagnostics from plain prompts
Tempus's PRISM2 foundation model matches or beats rivals on diagnostic and prognostic pathology tasks.
Stanford and Harvard launch MAST framework for tracking clinical AI progress
Stanford and Harvard researchers debut MAST v1.0, a new benchmark framework for measuring clinical AI performance across real healthcare tasks.
Frontier AI models beat specialized clinical tools in new benchmark
GPT-5.2, Gemini 3.1 Pro and Claude Opus 4.6 outperformed dedicated clinical AI systems in head-to-head testing across 1,100 medical questions.
AI model predicts immunotherapy response across cancer types
COMPASS, a pan-cancer foundation model, outperforms 22 methods across 16 clinical cohorts by analyzing tumor transcriptomes through immune concepts.
Michigan Medicine neuroimaging model outperforms GPT in brain scan analysis
Michigan Medicine researchers have introduced NeuroVFM, a visual foundation model trained on 5.24 million clinical brain scans that outperforms GPT-5…