Apple's AI model predicts health using wearable data
Apple and USC are working on an AI model that uses Apple Watch data—like sleep and activity—to predict health outcomes.
This is part of the Apple Heart and Movement Study, and while the findings are still waiting for peer review, it's a big step toward smarter health tracking.
Model trained on data from over 162,000 users
Their Wearable Behavior Model (WBM) looked at info from over 162,000 users (that's 2.5 billion hours of data!).
On 57 different health tasks, WBM beat standard methods in most cases, especially when paired with regular sensor data.
Still, there are some downsides: most participants were from the US, and pricey wearables might keep this tech out of reach for many.
Researchers say mixing behavioral data with sensor info could make future predictions even more accurate.