Apple Watch AI Model Enhances Health Predictions Using Behavioral Data

Apple developed a new AI model, the Wearable Behavior Model (WBM), trained on over 2.5 billion hours of Apple Watch data from more than 160,000 participants13.

Unlike traditional sensor-based methods that rely mainly on real-time biometric data, the WBM analyzes high-level behavioral metrics such as step count, sleep duration, heart rate variability, and mobility patterns135.

This approach enables the model to predict a broader range of health conditions, both static (e.g., medication effects) and transient (e.g., sleep quality, respiratory infections), with greater accuracy than previous methods134.

For pregnancy detection, the hybrid model (combining behavioral and biometric data) achieved up to 92% accuracy13.

The model was evaluated on 57 different health prediction tasks, focusing on identifying changes in behavior over days or weeks, making it especially effective for conditions that develop gradually13.

Apple gathered the training data through its Heart and Movement Study, with users voluntarily sharing data via their Apple Watch and iPhone13.

There are ongoing discussions about privacy implications of using such detailed behavioral data for health predictions5.

Sources:

1. https://www.phonearena.com/news/apple-watch-data-fed-an-ai-what-it-found-might-change-everything_id172176

3. https://www.macrumors.com/2025/07/11/apple-watch-ai-model-can-reveal-health-conditions/

4. https://applemagazine.com/apple-watchs-new-ai-model-uncovers-hidden-health-insights/

5. https://www.techradar.com/health-fitness/smartwatches/your-next-apple-watch-could-have-this-massive-ai-health-upgrade-and-no-new-sensors-are-needed