Apple Launches Health AI Study App With OpenEvidence, Enrolling Up to 200,000 U.S. iPhone Users

Apple is not yet following them into a product launch.
Apple is framing its OpenEvidence app as a study, not a chatbot launch like those from OpenAI and Anthropic.
Mark

Why does it matter that this is framed as a study rather than a product?

Mimi

Apple says a dedicated research study is the most thoughtful way to learn how people engage with, understand, and evaluate health AI. So the point is the learning, with up to 200,000 people over more than 12 months.

Mark

And the answers come from where, exactly?

Mimi

Clinical sources like NEJM, JAMA, and Nature, rather than a general pool of text. Apple says the app aims to simplify those answers for everyday questions.

Luke

Worth saying that grounding in those journals is Apple's description. The reporting doesn't say how well the answers perform, and nothing here tests that yet.

Mark

What about the personal data?

Mimi

Participants can share HealthKit data such as resting heart rate, sleep, activity, or lab results. Apple says it's anonymized before going to a secure third-party cloud provider and isn't used to train models.

Luke

And OpenEvidence deleting query and response data after 24 hours is also Apple's statement. We don't have an independent check on it.

Mark

Who can join?

Mimi

U.S. users over 21 with an iPhone 15 or later, iOS 26 or later, and Apple Intelligence on.

Luke

Also, 200,000 is a ceiling, "up to." We don't know how many will actually enroll, or when results will come out.

  • A crowded field of health chatbots has raised the stakes for who can be trusted with questions about the body.
  • Apple is testing a clinically sourced assistant against the flood of general-purpose AI answers people already rely on.
  • Participants must meet tight criteria, including U.S. residence, age over 21, recent iPhones, and Apple Intelligence enabled.
  • Users can share HealthKit data like heart rate, sleep, and lab results, a move that sharpens both usefulness and privacy concern.
  • Apple promises anonymized data, no AI training on user information, and 24-hour deletion of queries by OpenEvidence.
  • The study will run for over a year, and its findings may shape how consumer chatbots treat clinical grounding.

Apple has opened enrollment in a study that asks what happens when ordinary people bring their health worries to an AI whose answers are anchored in published medical research rather than the open sprawl of the internet. Built with the healthcare AI company OpenEvidence, the effort seeks up to 200,000 American iPhone users over more than a year, not to launch a product but to observe how people actually seek, understand, and weigh medical guidance. It arrives in a field already crowded with chatbots from larger labs, and it reflects a quiet recognition that trust in health information depends as much on where answers come from as on how fluently they are spoken.

Apple has opened enrollment in a study built around a simple but weighty question: what happens when ordinary people ask an AI about their health, and the answers are tied to serious medical literature? The work is being done through a dedicated app developed with OpenEvidence, a healthcare-focused AI company, and Apple hopes to enroll as many as 200,000 participants over more than twelve months.

The study arrives as OpenAI, Anthropic, and others have already pushed chatbots into health questions. Apple is not racing to join that launch. Instead it frames the project as research into how clinically grounded answers can actually serve people. To take part, a participant must live in the United States, be over 21, use an iPhone 15 or newer running iOS 26, and have Apple Intelligence turned on.

In the app, people can ask questions much as they would with a general chatbot, but the answers draw from sources such as The New England Journal of Medicine, JAMA, and Nature. Participants can rate responses, offer feedback, and answer monthly surveys. Those who choose to share HealthKit data, including heart rate, activity, sleep, or lab results, may receive more personal and specific suggestions. The app can also help people prepare for doctor visits and point them toward providers when a concern seems serious.

Apple has tried to address privacy directly. Data is anonymized before reaching a secure third-party cloud, will not be used to train AI models, and OpenEvidence says it deletes query and response data after 24 hours.

Sumbul Desai, Apple's vice president of Health and Fitness, described health as deeply personal and called a dedicated research study the most thoughtful way to learn how people engage with and evaluate health AI. The eventual findings, the company hopes, will inform the wider conversation about building tools people can trust, and may influence how clinical grounding is handled across consumer chatbots.

Apple has opened enrollment for a new research study that asks a plain question: what happens when ordinary people put their health questions to an AI whose answers are tied to clinical literature? The company is building the study around a dedicated app, made with the healthcare-focused AI company OpenEvidence, and it is aiming to enroll up to 200,000 participants for more than 12 months.

The timing places Apple alongside a crowded field. Big AI labs such as OpenAI and Anthropic have already made their bets on chatbots that field health-related queries, some with specialized features built for the purpose. Apple is not yet following them into a product launch. It is framing this as a study of how clinically grounded answers can help people.

Eligibility is narrow. Participants must be in the U.S. and over 21, using an iPhone 15 or later, running iOS 26 or later, with Apple Intelligence enabled.

Inside the app, users can ask health questions much as they would with ChatGPT or Claude. The difference lies in the sourcing: answers are grounded in clinical data from publications including The New England Journal of Medicine, the Journal of the American Medical Association, and Nature. Apple says the app will try to simplify its responses to everyday medical questions. Participants can rate the answers, give feedback on them, and take part in monthly surveys.

People who want more tailored responses can share data from HealthKit, such as resting heart rate, activity, sleep, or lab results, to get more personalized and pointed suggestions. Apple says the app can also help people prepare for a doctor's visit, and that it will direct users to specific providers when a question touches on a concerning medical issue.

Apple addressed privacy directly. Data will be anonymized before it is sent to a secure third-party cloud provider, and it will not be used to train any AI models. The company also said OpenEvidence deletes query and response data after 24 hours.

Sumbul Desai, M.D., Apple's vice president of Health and Fitness, said in a statement that responsible innovation starts with scientific rigor and clinical grounding. She called health deeply personal and described a dedicated research study as the most thoughtful way to understand how people engage with, comprehend, and evaluate health AI in everyday life. Insights from the study, she said, will help advance the broader scientific conversation and support the building of trusted tools.

What to watch is what the study eventually reports about how people use and judge these answers. Those findings could inform how clinical grounding is treated in consumer chatbots from OpenAI, Anthropic, and others.

Conducting a dedicated research study is the most thoughtful way to understand how people engage with, comprehend, and evaluate health AI in everyday life.
— Sumbul Desai, M.D., Apple vice president of Health and Fitness
Quieres la nota completa? Lee el original en TechCrunch ↗
Contáctanos FAQ