Apple Recruits Up to 200,000 U.S. Adults for Health AI Study With OpenEvidence

Measuring whether answers tied to clinical evidence are useful enough to act on
Apple's study tests clinically sourced AI answers rather than launching a general-purpose health chatbot.
Mark

Why does a study like this matter, when chatbots already answer health questions?

Mimi

Because Apple isn't just shipping an answerer. It's asking people to rate each response on accuracy, detail and relevance, and to fill out monthly surveys for more than a year.

Luke

Worth saying what that measures: how people perceive the answers. It's user ratings, not a check of clinical outcomes.

Mark

The answers link to journals like NEJM, JAMA and Nature. Is that the selling point?

Mimi

It's the center of the pitch. Apple's vice president of Health and Fitness, Sumbul Desai, talked about scientific rigor and clinical grounding.

Luke

That's Apple's framing. The source doesn't say how the app picks or summarizes studies, or how often it gets things wrong.

Mark

And the privacy side?

Mimi

Data is anonymized before reaching a third-party cloud, it won't train AI models, and OpenEvidence says it deletes queries after 24 hours.

Luke

All of that is the companies' own statement. Nothing here shows independent verification.

Mark

Does this feed into Siri or the Health app?

Mimi

Not now. The app doesn't use Apple Intelligence, and Apple hasn't said whether it will come later.

Luke

Right, so any talk of a product roadmap is speculation. What's confirmed is a study with a 200,000 ceiling, not that many enrolled yet.

  • Apple has opened enrollment for up to 200,000 U.S. adults, pulling everyday users into a clinical-grade AI experiment.
  • The stakes are rising as OpenAI and Anthropic already field hundreds of millions of health questions each week.
  • Participants rate every answer for accuracy, detail and relevance, while submitting monthly surveys to test real-world understanding.
  • The study is expected to run past 12 months, with results that could reshape Apple's health design and inform regulators and clinicians.
  • Key question still open: whether this answering capability will ever reach the Health app or Siri, which Apple has not addressed.

On a Monday in mid-November, Apple opened a door into a question people have been asking quietly for years: can a machine's answer to a worried question actually help a person understand their own body? Through its Apple Health AI Usage Study, built with the clinical AI platform OpenEvidence, the company is inviting as many as 200,000 American adults to test plain-language answers grounded in peer-reviewed medicine, rating each one and reporting back over a year or more. It is a deliberately patient experiment, arriving as the largest AI developers race to speak to people about their health, and its outcome will tell the wider world whether evidence-linked explanations can earn the trust that a confident chatbot has not yet fully earned.

On Monday, Apple opened enrollment for the Apple Health AI Usage Study and released a companion iOS app. The company is seeking as many as 200,000 participants across the United States, working with OpenEvidence, and expects the work to stretch well beyond a year.

To take part, a person must be 21 or older, live in the U.S., and own an iPhone 15 Pro or newer running iOS 26. Inside the app, they ask medical questions and receive plain-language answers drawn from journals such as The New England Journal of Medicine, JAMA and Nature. In return, participants rate each response for accuracy, detail and relevance, and complete monthly surveys. Apple's goal is to learn whether clinically sourced answers actually help people make sense of their health.

Those who choose can link HealthKit data, such as sleep, activity, resting heart rate or lab results, so answers reflect their own situation. The app can help prepare someone for a doctor's visit and points them toward specific providers when a question raises a concerning issue. Apple says data is anonymized before reaching a secure third-party cloud and will not train AI models. OpenEvidence says it deletes query and response data after 24 hours.

The app has firm limits. It does not run on Apple Intelligence, is not tied to Siri, and offers only standard iOS dictation for voice. Whether this answering capability will eventually reach Apple's Health app or Siri remains unanswered.

Apple's vice president of Health and Fitness, Sumbul Desai, M.D., framed the project around scientific rigor and clinical grounding, calling health deeply personal and a dedicated study the most thoughtful path to understanding how people judge health AI in daily life.

The effort comes as rivals move fast. OpenAI reported more than 300 million health queries weekly earlier this year. Apple's bet is different: rather than launch a general chatbot, it is measuring whether evidence-tied answers are clear and useful enough to act on. The findings could guide Apple's device design and give regulators and clinicians user-rated evidence, while offering OpenEvidence a large consumer test beyond its professional base.

On Monday, Apple opened enrollment for the Apple Health AI Usage Study and released a companion iOS app. The company is looking for as many as 200,000 participants in the United States. It is running the program with OpenEvidence, a clinical AI platform, and it expects the work to last longer than 12 months.

To join, a person must be an adult aged 21 or older and live in the U.S. They also need an iPhone 15 Pro or newer running iOS 26 or later. Inside the app, participants type in medical questions and get plain-language answers linked to peer-reviewed clinical literature. The material comes from publications including The New England Journal of Medicine, the Journal of the American Medical Association and Nature.

Participants have a job to do as well. They rate each response on accuracy, level of detail and relevance, and they are asked to complete monthly surveys. The aim, as Apple describes it, is to learn whether clinically sourced answers help people make sense of their health.

Enrollees can also choose to link records from HealthKit, such as activity, sleep, resting heart rate or lab results, so the answers reflect their own situation. Apple said the app can help someone prepare for a doctor's appointment. When a question raises a concerning medical issue, the app will direct the person to specific providers.

On privacy, Apple said data is anonymized before it reaches a secure third-party cloud provider and will not be used to train any AI models. OpenEvidence, whose platform is built for healthcare professionals, says it deletes query and response data after 24 hours.

The app has clear limits. It does not run on Apple Intelligence and is not connected to Siri. Dictation through the standard iOS keyboard is the only voice input. Apple has not said whether the underlying answering capability will eventually reach its Health app or Siri.

Sumbul Desai, M.D., Apple's vice president of Health and Fitness, said, "At Apple, we believe responsible innovation starts with scientific rigor and clinical grounding." She called health "deeply personal" and said a dedicated research study is the most thoughtful way to understand how people engage with, comprehend and evaluate health AI in daily life.

The study arrives as large AI developers push deeper into consumer health. OpenAI and Anthropic have both built specialized features for health questions, and OpenAI said earlier this year that users were submitting more than 300 million health-related queries each week. Apple's emphasis differs. Instead of shipping a general-purpose chatbot, it is measuring whether answers tied to clinical evidence are understandable and useful enough for people to act on.

The results could shape how health features are designed across Apple's devices, and they would give the company a body of user-rated evidence to cite as regulators and clinicians scrutinize medical AI. For OpenEvidence, a consumer study of this size is a large test bed beyond its professional base. What to watch is whether the answering capability moves into Health or Siri, something Apple has not addressed.

At Apple, we believe responsible innovation starts with scientific rigor and clinical grounding.
— Sumbul Desai, M.D., Apple vice president of Health and Fitness
Contact Us FAQ