Apple starts a health AI study for 200,000 iPhone users in the US
Apple and OpenEvidence will test for more than 12 months how users trust and understand health answers from AI.
On 15 November 2026, Apple started the Apple Health AI Usage Study in the United States. It works with OpenEvidence. This firm makes an AI tool that gives medical answers to health workers. Users of 21 years or more can start the study with an iPhone 15 Pro or a model after it. Apple tells that the study is to find the cause of trust in health AI.
How it works
Participants type usual health questions in a study app. Each answer has links to peer-reviewed sources. Participants can select to give HealthKit data. Examples are heart rate, sleep or lab results.
The numbers
The study is for more than 12 months. Participants rate answers and complete monthly surveys. Apple gave no results. Apple tells that the data has no user names, and that it does not use the data to train AI models. OpenEvidence removes questions and answers after 24 hours.
What is not known
Apple tells that tools that use large language models can give answers that are not full or not clear. The app gives no diagnosis and no treatment. Apple did not tell if a health tool will follow. It gave no test results for its safety guardrails.
Insight: Apple uses a partner model for a sensitive field
In China, Apple Intelligence will use the Qwen model of Alibaba. For health, Apple uses OpenEvidence. In the Siri AI preview, Apple gave developers no rules. This reads as a move of Apple to select specialist partners for sensitive fields and to keep the user. Leaders who make health AI must monitor which partner Apple names next.
Sources
Posted