Google Research moved symptom-checking AI beyond curated medical vignettes with a 13,917-participant national study. The result puts the emphasis on conversational history-taking, clinician comparison, and wearable biosignals rather than chatbot answer quality alone.
Google Research trained SensorFM on more than one trillion minutes of consented wearable data from five million people. The model beat feature-engineered baselines on 34 of 35 health prediction tasks, pointing to a more general route for wearable health AI.
OpenAI says GPT-5.5 Instant has pushed health responses close to its frontier Thinking models while reaching free ChatGPT users. The bigger signal is production data: flagged factuality issues in health answers fell 71% over two months.
Google Research is framing dermatology AI around user understanding, not just condition labels. A JAMA Dermatology study with 2,345 participants tested whether an AI-powered informational tool helped people identify skin concerns and choose better next steps.