Key Points
- Apple opened enrollment in a study asking up to 200,000 US users to rate AI health answers.
- The app is built with OpenEvidence, a clinical AI platform used by healthcare professionals.
- Apple says the ratings data will not be used to train AI models.
The latest:
Up to 200,000 US participants are being recruited to grade how accurate and useful AI-generated medical answers actually are, under a new Apple Health AI Usage Study that opened enrollment Monday. Apple is running it with OpenEvidence, a clinical AI platform used by healthcare professionals. Users submit health questions through a dedicated iOS app and score the responses they get back.
Details:
- How it works: Participants type health questions into the app and receive plain-language answers accompanied by links to peer-reviewed clinical sources. They then rate each response against criteria including accuracy, level of detail and relevance, according to Apple. The app accepts text prompts, with dictation available through the standard iOS keyboard.
- The AI partner: OpenEvidence generates its answers from medical literature including The New England Journal of Medicine, according to Apple. The platform is already used by healthcare professionals, which places the study’s test material closer to clinical reference material than to general-purpose chatbot output.
- Who can join: Enrollment is limited to people aged 21 and older based in the United States, Apple said. The hardware bar is narrow: an iPhone 15 Pro running iOS 26 or later is required, which rules out most of the installed base rather than opening the study to any recent iPhone owner.
- Not Apple Intelligence: Despite the subject matter, the app does not use Apple Intelligence and is not integrated with Siri, Apple said. The study is therefore measuring how people interact with an outside clinical AI system, not with Apple’s own consumer assistant stack.
- Personal health data: During enrollment, users can choose to share Apple HealthKit data such as activity, sleep or clinical lab results so that answers reflect their individual health context. Sharing is presented as optional rather than a condition of participation.
- Privacy terms: Apple says it has taken measures to keep participants’ information confidential, and states that data collected through the study will not be used to train AI models. The company did not detail the specific technical safeguards behind that commitment.
- The stated goal: Apple framed the project as an effort to better understand how people use conversational AI health tools and how useful the responses actually are, rather than as a product launch. No timeline for results or for any resulting feature was announced.
- Availability: The study app is available for download now, with enrollment beginning the same day the project was announced. Apple did not say when recruitment closes or whether the 200,000 figure is a hard ceiling.
- Also: Findings could shape how Apple deploys conversational health tools across its devices.
Between the lines:
The design points at a measurement problem rather than a product. By pairing an outside clinical platform with optional HealthKit data and a user rating system, Apple is collecting evidence on whether personalized AI health answers are judged trustworthy before putting such answers inside Siri. The explicit no-training pledge suggests Apple expects scrutiny of medical data handling.
What’s next
Watch whether enrollment reaches the 200,000 ceiling, when Apple publishes any findings, and whether health answering capability later appears inside Siri or Apple Intelligence rather than a standalone research app.