Apple has launched a research study that gives up to 200,000 people a health chatbot. It wants to understand what drives trust in the answers, how good those answers are, whether people understand them and what everyday health actions follow.
The Apple Health AI Usage Study is a virtual study with its own app, and it's launching in the United States, according to Apple's Newsroom. Apple built the study with OpenEvidence, which it calls a clinical AI platform "widely used by healthcare professionals". OpenEvidence's answers draw on sources such as The New England Journal of Medicine, the Journal of the American Medical Association and Nature.
Apple says the goal is to "inform the responsible design of health AI experiences for Apple and the industry".
What's in the app
- Everyday health questions. Apple says complex medical information is explained in "approachable language", and every answer links to peer-reviewed clinical citations.
- Optional HealthKit sharing. Participants can choose to share data such as resting heart rate, activity, sleep or lab results. Apple says this is to explore "how personal data integration impacts relevance and understanding".
- Ratings and surveys. TechCrunch reports that users can rate answers, give feedback and will be asked to fill in monthly surveys.
- Doctor prep. Apple says the study will look at whether health AI can help people prepare for appointments with their own doctors.
- Guardrails. The app is meant to detect "certain concerning questions" and suggest people follow up with a provider or other resources.
The fine print most people will miss
It's an iPhone 15 Pro, not an iPhone 15. Apple says participants need an Apple Intelligence-enabled iPhone 15 Pro or later running iOS 26 or later. TechCrunch's report says "iPhone 15 or later", but Apple's own wording names the Pro. The standard iPhone 15 doesn't support Apple Intelligence, so Apple's version is the one to go by.
Age and location. Participants must be 21 or older and in the US. Apple's announcement only covers the United States, with no word on other countries, Australia included.
It runs for a while. TechCrunch says Apple plans to run the study for more than 12 months.
What happens to your data. Apple says study data "will not be used to train AI models", and participants decide whether to share any HealthKit metrics or survey responses. According to TechCrunch, Apple also said the data is anonymised before it goes to a secure third-party cloud provider, and OpenEvidence deletes query and response data after 24 hours.
It's not a doctor. Apple says the app isn't meant to diagnose or treat any condition or replace medical advice, and that participants should check medical information with a doctor or clinician.
Why Apple's doing this
Apple says it plainly: consumer health tools powered by large language models "may at times offer responses that are incomplete or unclear". The study is meant to work out how to present health information more clearly.
"Health is deeply personal, and conducting a dedicated research study is the most thoughtful way to understand how people engage with, comprehend, and evaluate health AI in everyday life," said Sumbul Desai, Apple's vice president of Health and Fitness.
Robert M. Wachter, professor and chair of UCSF's Department of Medicine, pointed to the gap between the AI tools patients use and the ones doctors use. "Patients and clinicians are both turning to AI, but the tools they use are being built and tested in separate worlds," he said.
TechCrunch calls the potential "huge", noting that OpenAI said earlier this year that users ask it more than 300 million health-related questions a week. TechCrunch also says it's not clear whether Apple plans to bring AI-powered answers to the Health app or Siri.
Jax's take: If this ever reaches the Health app, testing it at this scale first is the right way round, and the 24-hour deletion rule matters more than any feature on the list. Whether this counts as science or just a very big beta depends on what Apple does with the results. If they're published openly, the "and the industry" line holds up.
Reporting notes



