AI

Apple Health AI Usage Study: what it tests and who can join

Key takeaways

  • The study is a research exercise, not a diagnostic product, and runs in a dedicated app built with OpenEvidence.
  • Participants can optionally share HealthKit data such as sleep, activity and lab results, and Apple says study data will not train AI models.
  • Enrolment is US-only, for people aged 21 and over with an iPhone 15 Pro or later on iOS 26 or later, so UK users cannot take part.

What happens when ordinary people put medical questions to a chatbot, and how much of the answer do they understand and trust? Apple has launched a US research study to find out. The Apple Health AI Usage Study, run with clinical AI platform OpenEvidence, is open to up to 200,000 participants and was announced on 15 November 2026.

What the Apple Health AI Usage Study actually is

It is a virtual research study delivered through a dedicated iOS app, not a feature of the Health app or Siri. Apple describes it as an effort to understand how people interact with conversational health AI tools, meaning chatbots that answer questions in plain language, usually powered by large language models (LLMs), the systems behind tools such as ChatGPT.

The aim is to examine what drives user trust, the quality of responses, comprehension and everyday health actions. The findings are meant to inform the design of health AI for Apple and, in Apple's words, the industry more widely.

The Verge's report adds that participants will rate the responses they receive on criteria including accuracy, detail and relevance. Participants are therefore both users and assessors of the responses.

How it works

OpenEvidence supplies the medical AI. Apple describes it as a clinical AI platform widely used by healthcare professionals that generates answers from established sources such as The New England Journal of Medicine, the Journal of the American Medical Association and Nature. Apple supplies the research experience around it.

Participants can ask the study app everyday health questions, and it returns answers in approachable language, with linked peer-reviewed clinical citations grounding each response.

Participants can also choose to share HealthKit data types, such as resting heart rate, activity, sleep or lab results, to see how personal data integration affects relevance and understanding. That choice is optional. The Verge notes it is offered at enrolment.

Two practical details matter. The app uses text prompts, and while iOS keyboard dictation works, The Verge reports it is not integrated with Siri or Apple Intelligence. Apple says it requires an Apple Intelligence-enabled iPhone 15 Pro or later running iOS 26 or later, although The Verge says the app itself does not use Apple Intelligence.

The study will also explore whether health AI can help participants prepare for visits with their own doctors.

Who it is for, and why Apple and OpenEvidence are doing it

Enrolment is limited to US residents aged 21 and over. The app is available to download now, and enrolment opened on launch day.

The rationale is a gap Apple itself acknowledges: consumer health tools powered by LLMs may at times give responses that are incomplete or unclear. Robert M. Wachter, professor and chair of the Department of Medicine at UCSF, framed the problem in Apple's announcement: patients and clinicians are both turning to AI, but the tools are built and tested in separate worlds. If both draw on versions of the same trustworthy tools, he argues, patients arrive prepared and conversations improve.

Sumbul Desai, Apple's vice president of Health and Fitness, said a dedicated study is "the most thoughtful way" to understand how people engage with, comprehend and evaluate health AI in everyday life.

What the limitations are

Apple is explicit that the study app is not meant to diagnose or treat any condition or replace the advice of medical professionals. Participants are told to consult a doctor or clinician about health concerns.

Safety guardrails are built in. The app is designed to detect certain concerning questions and suggest that the participant follow up with a provider or other resources. Apple says "certain", not all, so the guardrails are a backstop and not a guarantee.

There are structural limits too. The sample is restricted to people with a recent, high-end iPhone, which skews towards a particular slice of the population. The study is US-only. And it measures perception as well as accuracy: ratings of "accuracy" given by lay participants are not the same as clinical validation. The design reflects this, as it studies trust and comprehension alongside response quality.

On privacy, Apple says the study is designed with privacy protections, transparency and clinical guardrails. Study data will not be used to train AI models. Participants keep control over whether to share any HealthKit metrics or survey responses, and Apple says it has implemented security measures to protect the data it holds and the confidentiality of participants.

What this means in practice

For US iPhone owners who qualify, joining means contributing ratings that may shape how health AI is presented, in return for access to a citation-backed assistant. They should treat it like any health information source: useful for preparing questions, not for deciding treatment.

For UK readers, the study is out of reach. Nothing in the announcement gives a UK launch, date or availability. The findings could still matter here. Health data is special category personal data under UK GDPR, and the ICO publishes guidance on AI and data protection that any UK health AI deployment would need to consider. The ICO has not acted on this study, and the announcement does not say it has any UK dimension.

For IT and security teams in organisations that handle health data, the choice to make HealthKit sharing optional, and to state that study data will not train models, is a pattern worth noting when assessing vendors. The related question of who is accountable when health records meet AI is covered in this site's piece on Athenahealth's Claude link and HIPAA.

Where this is heading

Apple says insights from the study will help advance the broader scientific conversation. Watch for what the study publishes on trust, on comprehension, and on whether patients arrive at appointments better prepared. Those findings, more than the launch itself, will show whether conversational health AI earns a place in routine care.

Frequently asked questions

What is the Apple Health AI Usage Study?

A virtual research study and companion app, run with OpenEvidence, that examines how people use conversational health AI, including trust, comprehension and everyday health actions. It is not meant to diagnose or treat any condition.

Who can join the Apple Health AI Usage Study?

US residents aged 21 and over with an iPhone 15 Pro or later running iOS 26 or later. Apple is enrolling up to 200,000 participants.

Will Apple use my health data to train AI models?

Apple says study data will not be used to train AI models. Sharing HealthKit data, such as resting heart rate, sleep or lab results, is optional.

Can UK users take part in Apple's health AI study?

No. The study is launched in the United States and is open to participants across the US only.

More from Future Technology HQ