
Apple Recruits 200,000 iPhone Users to Test Health AI With OpenEvidence
Apple's Health AI Usage Study, launched Nov. 15 with OpenEvidence, seeks up to 200,000 US iPhone users to rate AI health answers with peer-reviewed citations over more than 12 months.
- By
- Nathan Brooks
- Filed
- Channel
- Consumer Tech
- Read
- 4 min read
Apple on Nov. 15 launched a virtual research study seeking up to 200,000 US iPhone users to measure how ordinary people judge AI-generated health answers. The study, built with clinical AI company OpenEvidence, is open to anyone aged 21 or older with an iPhone 15 Pro or later running iOS 26.
Apple announced the Apple Health AI Usage Study on stage at the HLTH conference in Las Vegas. It is research, not a product launch: Apple has not said whether AI-generated health answers will reach its Health app or Siri. The company wants to know what drives user trust, how well people comprehend the answers and what health actions they take afterward.
The participant bar is worth noting. A device requirement of iPhone 15 Pro or later trims the pool to Apple Intelligence-capable hardware, which conveniently matches the company's own platform ambitions.
How the app works
Participants type everyday health questions and receive answers in plain language with linked, peer-reviewed clinical citations. OpenEvidence's platform draws on sources including The New England Journal of Medicine and the Journal of the American Medical Association, combined with an Apple-designed research experience.
Users can volunteer HealthKit data such as resting heart rate, activity, sleep or lab results, letting the study test whether personal data makes answers more relevant. The study will also examine whether the tool helps people prepare for doctor visits.
Participants rate answers on criteria including accuracy, detail and relevance. TechCrunch reported that Apple plans to run the study for more than 12 months, that participants will complete monthly surveys, and that OpenEvidence deletes query and response data after 24 hours. The Verge reported that the app is text-based and has no integration with Siri or Apple Intelligence.
Apple has drawn the usual lines around it. The company says study data will not be used to train AI models, that participants control whether to share HealthKit metrics or survey responses, and that the app includes guardrails to detect concerning questions and point users to a provider. The app is not meant to diagnose or treat any condition.
"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.
What does OpenEvidence bring?
OpenEvidence sells a clinical answer engine to doctors. In January it said more than 40% of US physicians use the platform, disclosed alongside a $250 million Series D co-led by Thrive Capital and DST Global at a $12 billion valuation. That adoption figure is the asset Apple is renting.
OpenEvidence CEO Daniel Nadler framed the pairing at HLTH as bringing together the brand doctors trust most with the brand consumers trust most. It is a tidy sentence, and for once a marketing line that roughly matches the facts: one company owns the physician desktop, the other owns the consumer pocket.
Why the study matters commercially
The backdrop is volume. TechCrunch noted that OpenAI has said its users ask more than 300 million health-related questions a week. That traffic currently flows to a general-purpose chatbot with no clinical citation discipline attached. Apple's study is, among other things, an attempt to define what a credible, evidence-linked answer should look like before competitors set the default.
There is also regulatory positioning. By running a formal research study with consent, deletion windows and explicit non-training commitments, Apple builds a record it can hand to regulators and clinicians later. A 200,000-person cohort tracked for more than a year is a serious dataset by health research standards, where studies of a few thousand participants are routine.
Apple has also kept the study deliberately narrow. The text-only app, the 24-hour deletion of queries, and the refusal to fold the tool into Siri all reduce the liability surface while the company learns how badly things can go. The guardrails that direct concerning questions to a provider suggest Apple already knows the failure mode.
Separately, Apple this month extended its redesigned Health app, which it says is built with Apple Intelligence at its core, to more languages including Chinese, French, Japanese and Spanish. The Health app's international rollout and the AI study are separate efforts, but they point the same direction: Apple is assembling evidence and infrastructure before it commits health answers to its mainstream products.
What happens next
The study runs for at least a year, with findings that Apple says will inform the design of health AI tools for Apple and the wider industry. Whether that translates into a Siri that cites the New England Journal of Medicine remains unstated, and Apple is careful to keep it that way.
Source: mlq.ai
More from Nathan Brooks
Show full bio
Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.
1,010 articles
Related articles
apple-recruits-up-to-200-000-iphone-users-for-health-ai-study
Apple Recruits Up to 200,000 iPhone Users for Health AI Study
apple-tests-four-daily-ai-caps-and-tighter-mac-agent-permissions
Apple tests four daily AI caps and tighter Mac agent permissions
mercedes-brings-hands-free-highway-driving-to-the-us-at-up-to-85-mph
Mercedes brings hands-free highway driving to the US at up to 85 mph in 2027
iphone-duo-gains-camera-face-unlock-in-ios-27-3-beta-with-strings-attached
iPhone Duo Gains Camera Face Unlock in iOS 27.3 Beta — With Strings Attached



