Apple opens health AI study to up to 200,000 US iPhone users
Edited and reviewed by AI Ping Editorial Team
3-line summary
- Apple began recruiting US adults aged 21 and older for its health AI study on Nov. 15.
- The free English-language study app requires an Apple Intelligence-compatible iPhone running iOS 26 or later.
- Participants can ask an OpenEvidence-powered chatbot health questions and choose whether to share health records for personalized answers.
Apple began recruiting for its Health AI Usage Study in the United States on Nov. 15. The company plans to accept up to 200,000 eligible adults aged 21 and older. They can choose to enroll through a free iPhone app designed by Apple and powered by technology from OpenEvidence. The study will examine how people understand and judge the answers they receive when they ask a chatbot about their health.
The app requires an iPhone that supports Apple Intelligence and runs iOS 26 or later. Within the iPhone 15 lineup, the 15 Pro and 15 Pro Max meet the hardware requirement. The US App Store lists English as the study app’s language and marks it 18+ for content. Apple sets the minimum age for enrolling in the research at 21.
Participants choose which health records to share
Participants can ask everyday questions and separately decide whether to let the app use information stored in HealthKit. Apple gives resting heart rate, activity, sleep and lab results as examples of data people can choose to provide. HealthKit brings together health records from the iPhone and Apple Watch. Apple’s platform security guide says users can review and revoke access by data type in iPhone settings. The App Store listing says the app can use the records a participant shares to personalize its replies. The researchers will look at whether that makes the replies more relevant or easier to understand.
Apple says it will keep study data out of AI training. The company told ABC News that it designed the system to strip identifying information from questions before passing them to OpenEvidence, which will delete question-and-response data within 24 hours. Apple’s App Store privacy label separately says a user ID may be linked to a participant’s identity for analytics or app functionality. It lists performance diagnostics under data that may be collected without being linked to the user.
Participants review the study information and give consent before enrolling. The general health studies privacy policy linked from the listing says participants can withdraw at any time. The researchers then stop receiving new study data about them and may keep or continue using information already collected. Deleting the app alone leaves enrollment in place. Each study’s consent document sets out its specific rules for retention and withdrawal. Apple says the study-specific terms govern if they differ from the general policy.
The app links to research behind its answers
The chatbot presents medical information in plain language and includes links to peer-reviewed research. OpenEvidence also runs a separate clinical app that requires verification as a US healthcare professional. In Apple’s study app, participants can rate answers and choose whether to respond to surveys. Apple will also study whether AI can help people prepare for appointments with their own doctors.
Apple describes the app as a source of information and excludes medical advice, diagnosis and treatment from its scope. It says the app is designed to recognize certain concerning questions and may direct participants to a healthcare provider or other resources. The launch announcement describes the study’s aims and enrollment. It gives no findings on how accurately the app answers health questions or how participants act on the information.
A Pew Research Center survey of 3,488 US adults conducted June 22 to 28 found that 34% had used an AI chatbot for at least one of eight health-related purposes. Some 22% of US adults said they had used one to learn more about a doctor’s diagnosis, and 20% had done so to understand lab results. All three figures refer to US adults overall.
Pew also asked people who had used a chatbot for health purposes how comfortable they would be sharing personal health information with one. Twenty-nine percent said they would be very or extremely comfortable, and 26% said they would be not too comfortable or not comfortable at all.
Apple’s separate announcement earlier in November covered additional languages, including Korean, for the regular Health app. The Nov. 15 research launch specifies US enrollment, and the study has its own App Store listing. Apple’s launch notice gives no deadline to sign up or date for releasing the research findings.
Sources
- 건강 AI 연구 시작 발표 · Health AI study announcement — Apple
- 애플 건강 AI 연구 앱 · Apple Health AI Study App — Apple App Store
- 애플 인텔리전스 지원 기기 · Apple Intelligence device requirements — Apple Support
- 헬스킷 접근 권한 · HealthKit data permissions — Apple Support
- 건강 연구 개인정보 처리방침 · Health Studies Privacy Policy — Apple
- 애플 연구 앱 보도와 데이터 처리 관련 회사 설명 · Health chatbot reporting and Apple’s data handling statement — ABC News
- 의료인용 오픈에비던스 앱 · OpenEvidence app for clinicians — OpenEvidence
- 미국 성인의 건강 챗봇 이용 설문 · Survey of US health chatbot use — Pew Research Center
- 기본 건강 앱의 언어 지원 확대 · Language expansion for the regular Health app — Apple