Apple delves into artificial intelligence applied to healthThe Cupertino multinational is developing a new generation of improvements for the Apple Watch, focused on extracting and interpreting behavior patterns based on the data collected by the smartwatch. It's no longer just about measuring pulse or steps, but about using AI to decipher how each person's daily routines evolve and identify signs of health changes or issues even before symptoms appear.
The approach goes a step beyond traditional sensors and has already been tested in one of Apple's most ambitious health studies, the Apple Heart and Movement StudyThis research has served as the basis for the AI model, trained with anonymous and voluntary information collected during more than 2.500 billion hours, from more than 160.000 Apple Watch and iPhone users.
How does AI work on the Apple Watch?
The key lies in the Wearable Behavior Model (WBM), the new AI model presented by Apple. This system abandons the processing of raw data from sensors (such as the photoplethysmograph or the accelerometer) and focuses on 27 interpretable metrics that summarize daily behaviors: active energy, gait stability, sleep duration, resting heart rate, respiratory rate, among others.
These indicators allow AI detect variations medium and long term, such as changes in sleep quality, mobility, physical activity or even the onset of transient states such as pregnancy. In addition, the model has achieved up to 92% accuracy in identifying pregnancies, surpassing classic methods and without resorting to specific menstrual or temperature monitoring sensors.
The advantages of this approach are also reflected in the ability to anticipate the start of treatment with certain medications, such as beta-blockers, to closely monitor recovery from an injury or to detect cognitive impairment through changes in daily activity.
Improving prediction with the combination of AI and traditional sensors
While the WBM model excels at interpreting behaviors and anticipating states, there are conditions for which direct physiological data remains dominant. For example, the diagnosis of diabetes or monitoring certain medications works best with pure photoplethysmography (PPG) measurements.
La most effective strategy It is the combination of both sources: the analysis of behavioral patterns through AI and real-time data captured by sensors. By merging both models, Apple has managed to considerable improvements in accuracy for most of the tasks analyzed, confirming that the collaboration between technology and algorithm can offer much more reliable early diagnoses.
Users do not need to have a cutting-edge device: these capabilities could be distributed via software updates. watchOS and the Apple Intelligence Platform, leveraging the infrastructure already in place in watches launched in recent years.
Privacy and the Future of Digital Health on the Apple Watch
The use of artificial intelligence to interpret personal health data has also raised concerns. concerns about privacy and control of biometric informationApple claims that the model's development is based on anonymized data and voluntary user participation in research programs. However, both access to these advanced systems and protection against improper or commercial use of information remain under debate.
Another aspect that has yet to be officially confirmed is whether these features will be part of the standard offering of the devices or if some advanced features will require additional payments, following the example of other brands in the sector.
For now, the integration of AI in the health field focuses on a a more global and preventive vision of well-being. It allows the Apple Watch to not only record data, but also to interpret it, anticipate risks, and suggest individualized actions to improve quality of life.
