Imagine if one night’s sleep could tell doctors more about your long-term health than your last check-up. That’s exactly what scientists have achieved with a new artificial intelligence model published in Nature Medicine. Researchers at Stanford Medicine trained an AI system on over 585,000 hours of detailed sleep recordings — including brain activity, heart rhythms, breathing patterns, and muscle signals — from thousands of participants. By “learning the language of sleep,” this model can now predict the risk of developing 130 different diseases from just one night’s sleep data.

From serious conditions like dementia, heart failure, stroke, chronic kidney disease, and cancer to overall mortality risk, the AI’s predictions showed strong accuracy — often years before clinical symptoms appear.

This research highlights sleep not just as rest, but as a rich well of early-warning health signals. While this technology isn’t yet part of routine care, it points to a future where your sleep patterns could one day help guide personalized prevention and early intervention.