Apple's Health AI Study: Who Can Join and What Data It Collects
Apple and OpenEvidence are recruiting up to 200,000 US iPhone owners to test how people trust, understand and act on AI health answers.
Apple launched a virtual research study on Nov. 15, 2026, with an app that supplies AI health answers from partner OpenEvidence. Adults 21 and older in the US with a recent iPhone can apply, and sharing Apple Health data is optional. Apple's privacy promises have not been independently audited, and key design details are still unpublished.
Apple and the clinical AI company OpenEvidence opened a virtual research study on Nov. 15, 2026, built around a dedicated app in which participants ask health questions and judge the AI's answers. According to Apple's announcement, up to 200,000 eligible people age 21 and older across the United States can take part. The company says the study will look at what drives trust in health AI, the quality of the responses, how well people understand them and what they do with them in everyday life.
The study arrives as AI chatbots are already a routine source of health information for many Americans. It is a research project, not a product launch: Apple has not said whether AI-generated health answers will ever appear in its Health app or in Siri. The practical questions for readers are narrower and more immediate, namely who qualifies, what the app collects, and how much of Apple's privacy account can be checked.
What the study tests, and who runs it
Apple announced the project on Nov. 15, 2026, describing it as a virtual research study with an associated app, run in collaboration with OpenEvidence. Digital Health News reports a division of labor: Apple designs the research experience, while OpenEvidence supplies the clinical AI that produces the answers.
The stated aims, as relayed by PYMNTS from Apple's release, are to examine user trust, response quality, comprehension and everyday health actions. MobiHealthNews adds that Apple also wants to explore how health AI could help participants prepare for visits with their own doctors.
TechCrunch quotes Sumbul Desai, Apple's vice president of Health and Fitness, as saying that a dedicated research study is the most thoughtful way to understand how people engage with, comprehend and evaluate health AI. Apple also acknowledged, per news reports quoting the company, that consumer health technology powered by large language models may at times give answers that are incomplete or unclear.
What it is not
The study should not be confused with the separate Apple Health Study in Apple's Research app, which covers many health topics. According to the Apple Heart and Movement Study site run by Harvard's Brigham and Women's Hospital, that study does not require an Apple Watch. The AI study has its own dedicated app, though that point rests on a single secondary-source claim.
The app also does not offer medical care. Reports say it cannot provide medical advice, diagnosis or treatment, and that it can detect concerning questions and suggest working with a healthcare provider. Engadget reports that Apple says the app is not integrated with Siri or Apple Intelligence.
Who can join
Apple's release sets the core rules: eligible participants must be 21 or older and live in the United States, and enrollment is capped at 200,000. Multiple outlets report that the phone must run iOS 26 or later. TechCrunch and Startup Fortune add that Apple Intelligence must be switched on.
| Item | What has been reported | Reported by |
|---|---|---|
| Maximum enrollment | 200,000 | Apple Newsroom |
| Minimum age | 21 | Apple Newsroom |
| Geography | United States | Apple Newsroom |
| Software | iOS 26 or later | Multiple outlets |
| Device | iPhone 15 Pro or later (TechCrunch says iPhone 15 or later) | Wareable, MobiHealthNews, TechCrunch |
| Duration | More than 12 months (not confirmed by other outlets) | TechCrunch |
| Deletion of questions and responses by OpenEvidence | Within 24 hours, per Apple | ABC News, TechCrunch |
The device question
Reports disagree on the exact iPhone. Wareable, MobiHealthNews, Neowin and iGeeksBlog say an Apple Intelligence-enabled iPhone, meaning the iPhone 15 Pro or later, while TechCrunch says iPhone 15 or later with Apple Intelligence enabled. Only the Pro models of the iPhone 15 generation support Apple Intelligence, so "iPhone 15 Pro or later" is the more likely wording, but that could not be checked against Apple's own page.
Cost, credentials and length
iGeeksBlog reports that the app, called Apple Health AI Study App, is free on the US App Store. No professional credentials are required. That is a notable change from OpenEvidence's regular app, whose App Store listing says it is available only to health care professionals.
Two details remain unsettled. TechCrunch reports that the study will run for more than 12 months, but iGeeksBlog and Data Studios say Apple has not published a duration. Data Studios also says Apple has not published any compensation.
The 200,000 figure is a ceiling, not a headcount. No enrollment numbers have been reported since the launch.
What participants do and what data is collected
ABC News reports that participants ask everyday health questions and receive answers with linked, peer-reviewed clinical citations. TechCrunch says the answers are grounded in sources that include The New England Journal of Medicine, JAMA and Nature. Startup Fortune, relaying TechCrunch, says participants rate each response on qualities such as accuracy, detail and relevance, and complete a monthly survey.
KRTV lists the example uses Apple has offered:
- Understanding symptoms or conditions
- Putting personal health data in context
- Preparing for appointments
- Creating exercise and meal plans
The data that is clearly part of the study
No official data inventory has surfaced in the coverage reviewed. From how the app works, the study evidently handles the questions participants type, the AI's responses, the ratings and the monthly survey answers. Neowin reports that participants control whether to share their survey responses.
The optional layer: Apple Health data
HealthKit is the data store behind the iPhone Health app, holding measures such as sleep, activity, resting heart rate and lab results. According to MobiHealthNews, participants will be invited to share metrics such as heart rate, activity, sleep and lab results. Apple's stated purpose is to explore how integrating personal data affects the relevance and understanding of answers.
Sharing is not required. TechTimes answers "No" to whether HealthKit access is needed, and multiple other sources describe it as optional. The exact list of data types is less certain than the optional status itself.
What coverage does not say
Coverage does not say whether device identifiers, location, demographics or usage logs are collected, and none of that should be assumed either way. One blog claims medical records enter the study only if a person has imported them into the Health app and separately consents, but that rests on a single weak source and could not be confirmed against Apple.
Privacy promises, and who has checked them
Apple's release, as relayed by multiple outlets, makes three main commitments. Study data will not be used to train AI models. Questions pass through protections designed to strip identifying information before they reach OpenEvidence. And, per ABC News and TechCrunch, OpenEvidence will delete question-and-response data within 24 hours and will not train its model on the queries.
Apple has also said, according to a report relayed by the Pebblous blog, that data will be anonymized before being sent to a secure third-party cloud provider. That provider has not been named in the coverage reviewed.
Repetition is not verification
Most coverage reproduces the same Apple release, so agreement among outlets reflects repetition rather than independent confirmation. CASRAI makes the point directly: these are attributed descriptions of the study's safeguards, not the result of an independent privacy audit. The full path a question takes between the iPhone, Apple, the cloud provider and OpenEvidence has not been publicly described.
Reports also conflict on retention. The 24-hour deletion appears in outlets citing Apple, yet Data Studios says retention periods have not been stated. Apple's consent form and privacy notice, which would settle such points, were not available for review.
There is also an open tension. If OpenEvidence deletes queries within a day, the lasting research record would presumably be the ratings, surveys and any HealthKit data. No source explains what is retained for analysis, or whether "study data" in Apple's pledge includes the question-and-answer text that Apple itself holds.
Why a health AI study, and why now
A KFF tracking poll from March 2026 found that about a third of US adults (32%) had used AI chatbots for health information in the past year; Becker's Hospital Review reports the poll surveyed 1,343 adults. Among those users, 41% had uploaded personal medical information, which equals 13% of the public. Adults under 30 were about three times as likely as those 50 and older to use AI for mental health information, at 28% versus 8%.
Trust lags use. KFF found that 67% of adults trust AI tools "not too much" or "not at all" for reliable health information. The poll is a self-report survey with sampling error. An earlier KFF measure from June 2024 found 17% using chatbots at least monthly for health information, but it asked a different question, so the two numbers are not directly comparable.
A crowded field
TechCrunch reports that users were asking OpenAI's ChatGPT more than 300 million health-related queries each week, while KFF, citing OpenAI, says more than 40 million people worldwide turn to ChatGPT daily for health information. Both are relayed figures; OpenAI's own statement was not located. The Next Web reports that OpenAI launched ChatGPT Health in January, with Microsoft and Perplexity following in March.
What is known about OpenEvidence
OpenEvidence is a clinical AI platform built for health professionals. A JAMA cross-sectional study published Oct. 15, 2026, analyzed 271 million queries from 952,008 verified US clinicians between January 2024 and June 2026, according to a summary on distilinfo. Monthly active users rose from 355 in January 2024 to 703,265 in June 2026.
The same summary reports that three co-authors, including OpenEvidence's chief medical officer, were company employees and that the company supplied the data and contributed to the analysis. Another source says the company did not fund the research. The paper measured usage, not accuracy or patient outcomes, and clinicians' query habits do not show how laypeople will fare with a tool designed for doctors.
eMarketer reports that OpenEvidence keeps its platform free for physicians by selling targeted digital ads to pharmaceutical and medical device companies. Whether ads appear in the consumer study app is unknown. OpenEvidence's own App Store text calls it an official AI partner of NEJM and JAMA, which is a vendor claim.
What the study design leaves open
The coverage reviewed turned up no published protocol, registry entry, ethics (IRB) approval, named principal investigator or outcome definitions. Nothing explains how trust, comprehension or quality will be measured, or who will rate quality. Participant self-ratings differ from expert review, and it is unclear whether any comparison group, such as other chatbots, is planned.
Apple's earlier Research-app work offers context but not evidence about this study. Its Heart and Movement Study produced peer-reviewed output, with data from over 30,000 participants analyzed. No commitment to publish results from the AI study has been found.
Who ends up in the sample
Participants need a recent iPhone, iOS 26, Apple Intelligence turned on and an interest in health AI. Chief Healthcare Executive flags that these requirements could bias who takes part and who has access, though no demographic data on participants exists yet. KFF finds higher AI health use among younger, lower-income, Black and Hispanic adults. They may be less likely to own newer iPhones, but that is an inference rather than a finding from any source.
Independence
OpenEvidence both supplies the AI and collaborates on the study, while Apple controls the research. That is a potential conflict of interest for any finding about answer quality, and no independent evaluator has been named. Accuracy claims about OpenEvidence are largely the vendor's own. An arXiv preprint that has not been peer reviewed compared clinical AI tools on real point-of-care queries; it says OpenEvidence co-developed the data collection plan and paid survey responders, while none of the authors is affiliated with the company.
What to watch
The study speaks to a real gap. Many Americans already ask chatbots about their health, yet most say they trust the tools little, and little is known about how everyday users read, rate and act on the answers. A large, structured look could help, if its methods are open to scrutiny.
The signals worth following are concrete: whether Apple publishes its consent and privacy documents, names an independent evaluator, reports who actually enrolled, and commits to releasing results. Until then, the study is best read as a recruitment drive with stated safeguards, not as evidence that AI health answers are reliable.
Frequently asked questions
Who can join Apple's Health AI Usage Study?
According to Apple, eligible participants are age 21 or older and live in the United States, and enrollment is capped at 200,000. Outlets report the phone must run iOS 26 or later with Apple Intelligence turned on. Most reports specify an iPhone 15 Pro or later, though TechCrunch says iPhone 15 or later.
Is Apple's AI health study free to join?
iGeeksBlog reports that the Apple Health AI Study App is free on the US App Store. Data Studios says Apple has not published any compensation for participants. No medical credentials are needed, unlike OpenEvidence's regular app.
Do participants have to share Apple Health data?
No. Multiple sources describe sharing HealthKit data, such as heart rate, activity, sleep and lab results, as optional. Apple says participants will be invited to share it to explore how personal data affects the relevance and understanding of answers.
Will Apple or OpenEvidence use study data to train AI?
Apple says study data will not be used to train AI models, and that OpenEvidence will not train its model on the queries and will delete question-and-response data within 24 hours. These are Apple's statements, relayed by news outlets. No independent audit has been reported.
Will this bring AI health answers to Siri or the Health app?
Apple has not said. The study is research, and Engadget reports that Apple says the app is not integrated with Siri or Apple Intelligence. The app also cannot provide medical advice, diagnosis or treatment.
Sources (27)
- Apple launches research study to explore the use of AI for health , Apple Newsroom (2026-11-15)
- Apple launches new app to study how people use AI for health , TechCrunch (2026-11-15)
- Apple launches an AI health tool as part of a research study , ABC News
- Apple launches research study with OpenEvidence , MobiHealthNews
- Apple's new research app is designed to make AI health tools more user-friendly , Engadget
- Apple and OpenEvidence Seek 200,000 Consumers for Health AI Study , PYMNTS
- Apple's new app aims to study how people use AI for health , Neowin
- Apple explores conversational health AI with new 200,000-user research study , Wareable
- Apple's AI health study is open to up to 200,000 U.S. iPhone users , iGeeksBlog
- Apple Health AI study app: 200,000 U.S. iPhone users, OpenEvidence, eligibility , Data Studios
- Apple Launches AI Health Study With OpenEvidence , Digital Health News
- Apple launches health AI study with app that answers medical questions , KRTV
- Apple Launches Health AI Usage Study to Explore How People Use AI For Medical Advice , Tech Times (2026-11-16)
- Apple launches a study app with OpenEvidence to see how people use AI for health , Startup Fortune
- Apple Health AI Usage Study: What It Measures , Pebblous
- Apple launches AI health research study , CASRAI
- Apple plans to study how consumers use AI for health needs: Questions and answers , Chief Healthcare Executive
- Apple and OpenEvidence test whether people trust AI with health questions , The Next Web
- Poll: 1 in 3 Adults Are Turning to AI Chatbots for Health Information, Equaling the Share Who Use Social Media for Health , KFF
- Poll: Most Who Use Artificial Intelligence Doubt AI Chatbots Provide Accurate Health Information , KFF
- The Growing Use of Artificial Intelligence in Health Care and Implications for Disparities , KFF
- 32% of adults turn to AI for health advice: KFF , Becker's Hospital Review
- OpenEvidence Study: Clinicians Use AI Most for Treatment 2027 , distilinfo (2026-10-16)
- Physicians' AI search patterns give pharma new ways to deliver more relevant advertising , eMarketer
- Expert Evaluation of Clinical AI Tools on Real Point-of-Care Clinical Queries , arXiv (preprint, not peer reviewed)
- Apple Heart and Movement Study , Brigham and Women's Hospital
- OpenEvidence (App Store listing) , Apple App Store