Google Research AI detects hypertension signs from smartwatch sensors

Google Research, working with Beth Israel Deaconess Medical Center and Harvard Medical School, published a Nature study describing an AI system that uses standard consumer smartwatch sensors to flag signs of high blood pressure without any cuff calibration. The model, called AI-PPG-ACC-HTN, learns from pulse and motion waveforms gathered over multiple days and combines the aggregated result with the wearer's height to produce a hypertension probability score. In a 196-participant study across five US sites, benchmarked against 24-hour ambulatory blood pressure monitoring, the researchers report accuracy comparable to conventional cuff-based screening workflows. The authors describe the work as early feasibility research rather than a product, and say the findings informed the blood pressure trends Health Guardian feature in the Google Health app.