Apple’s new health AI app is a study, not your doctor

Someone checking the heart health settings on their Apple Watch
(Image credit: Apple)

Apple’s new health AI app lets study participants ask medical questions, but it isn’t a substitute for a clinician. The US research project will examine how people understand, use, and trust its answers.

The short version

  • Apple announced the Apple Health AI Usage Study on November 15, working with OpenEvidence. Participants can ask everyday health questions and see linked, peer-reviewed clinical citations supporting responses.
  • The research goes beyond whether an answer sounds convincing. Apple says it will examine response quality, comprehension, trust, and everyday health actions, including whether the app can help people prepare for medical appointments.
  • Sharing HealthKit information, such as sleep, resting heart rate, or lab results, is optional. Apple says participants also control whether to share survey responses, and that study data won’t be used to train AI models.
  • Apple acknowledges that responses can be incomplete or unclear. Its App Store listing explicitly says neither the app nor the study provides medical advice, diagnosis, or treatment.

Does this affect you?

  • US adults considering enrollment: The study is open to up to 200,000 eligible participants age 21 and older with an Apple Intelligence-enabled iPhone running iOS 26 or later. The app is free to download; its 18+ App Store content rating doesn’t lower the study’s enrollment age.
  • People connecting Apple Health data: You can choose whether to share HealthKit metrics for the study to explore more personalized answers. That sharing isn’t compulsory participation in AI-model training; Apple expressly rules out using study data for that purpose.
  • Anyone with a health concern: Apple says to consult a doctor or clinician to address concerns or verify medical information. Clinical citations don’t change the app’s stated medical limits.

Editor's take

iMore Editor's take

If any tech company can convince people to share their health and fitness data with AI, it’s Apple. But trusting Apple to protect that information isn’t the same as trusting an AI app to give you a reliable answer.

OpenAI reports that over 40 million people use ChatGPT every day for health questions, and King's College London says 1 in 7 people have used AI instead of seeing a GP. With so many people turning to AI to ask for medical advice, Apple is looking to understand why and how. Apple says the goal of the study is to "examine what drives user trust, the quality of responses, comprehension, and everyday health actions; the study aims to inform the responsible design of health AI experiences for Apple and the industry."

I can see the appeal of helping Apple understand how people use AI for health questions, but I'd urge those who do this regularly to err on the side of caution. This is for research, and absolutely not a finished product, so please treat participation as a contribution to the cause, rather than some early access to your future AI doctor.

Worth reading

  • Apple: The study’s aims, eligibility requirements, and stated privacy safeguards.
  • TechRadar: An opinion piece on the privacy and trust questions raised by sharing medical data with ChatGPT Health.
  • Tom’s Guide: A first-person account of an incorrect pregnancy assessment from ChatGPT, illustrating why confident AI answers aren’t necessarily reliable.
Why Trust iMore

We used our AI-assisted editorial tool to help create this iMore explainer, summarizing key facts, outlining what they could mean for readers, and highlighting unanswered questions. iMore's Editor-in-Chief wrote the “Editor’s take” without AI assistance and reviewed and edited the full article before publication. Our editors at Future PLC retain responsibility for the final article and the decision to publish.

iMore Labs is a team exploring how AI-assisted tools can support journalism, from organizing source material and summarizing key facts to simplifying repetitive editorial tasks. We build and test workflows designed to give journalists more time for reporting, analysis, and original ideas, with the aim of making useful tools adaptable across publications and subject areas.

Our work keeps human judgment at the center: journalists choose the stories, shape the coverage, write the opinions, and retain responsibility for what gets published.