An Apple-backed study has shown that AI analyzing Apple Watch and iPhone behavior data can predict pregnancy with 92% accuracy, offering a glimpse into the future of wearable-based health diagnostics. (Source: Image by RR)

AI Study Combines Biometric and Behavioral Data for Accurate Pregnancy Prediction

A new Apple-supported study suggests that Apple Watches and iPhones could one day be used to detect early pregnancy with surprising accuracy. The research, leveraging both biometric and behavioral data collected from wearables, found that an AI model could predict pregnancy with 92% accuracy—without requiring a traditional test. Instead of just tracking basic health stats like heart rate or temperature, the model also factored in behavioral shifts such as sleep patterns, walking gait, and physical activity levels, which tend to change significantly in early pregnancy.

The study, as noted in pcmag.com, presents a compelling case for the power of combining physiological and behavioral data in health prediction models. Pregnancy was just one condition analyzed: the AI also predicted diabetes with 82% accuracy, infections at 76%, and injuries at 69%. These results demonstrate the potential for consumer-grade wearables to be repurposed as health screening tools, expanding far beyond their original fitness tracking capabilities. Though not peer-reviewed yet, the findings highlight the growing role of AI in preventive medicine.

The AI model used in the study—called the Wearable Health Behavior Foundation Model (WBM)—was trained on massive datasets from the Apple Heart and Movement Study. This included more than 15 billion data points gathered from 162,000 Apple device users. Specifically for the pregnancy prediction test, the researchers analyzed data from 430 pregnancies among 385 participants, and compared it with 25,225 non-pregnant control participants to assess accuracy.

While Apple has yet to formally announce any commercial implementation of the study’s findings, the implications are notable. As the U.S. government expresses interest in expanding wearable use across the population, these findings could help shape future health policy and innovation. The technology points to a future where wearables not only track fitness, but also detect major health events—before symptoms even surface.

read more at pcmag.com