Apple launched the Apple Health AI Usage Study and a dedicated mobile application on Sunday, November 15, 2026, aiming to recruit up to 200,000 participants across the United States to examine how consumers interact with conversational artificial intelligence for medical information. The virtual research initiative, conducted in partnership with clinical AI platform OpenEvidence, will evaluate whether consumers understand and trust AI-generated health advice, how they integrate personal biometric data into their queries, and whether these tools help patients prepare for appointments with their physicians.

Roughly one in three U.S. adults already rely on AI chatbots for health information, a trend that is even more pronounced among younger adults, according to a KFF survey cited by Fierce Healthcare.

While major AI laboratories such as OpenAI and Anthropic have already introduced general chatbots capable of answering medical queries—with OpenAI reporting earlier this year that users submit over 300 million health-related questions each week—Apple is taking an observational approach before deciding how to integrate conversational health tools into its wider ecosystem.

Bridging Clinical Literature and Consumer Queries

Rather than relying on open-web scraping, the Apple Health AI Study app utilizes underlying clinical technology from OpenEvidence, a medical search engine and decision-support tool built for healthcare professionals. According to Apple Newsroom, the platform generates responses by drawing directly from licensed, peer-reviewed medical literature and clinical guidelines through formal partnerships with The New England Journal of Medicine, the Journal of the American Medical Association (JAMA), Nature, the Cochrane Database of Systematic Reviews, and the National Comprehensive Cancer Network.

Within the standalone study app—which operates separately from the standard Apple Health and Apple Research apps—participants can submit text-based questions regarding general wellness, specific physical symptoms, fitness routines, and recovery from medical procedures.

The system translates complex medical findings into accessible, everyday language while providing links to the underlying peer-reviewed citations. Users are asked to rate the quality of the AI's responses, indicate what information was useful or lacking, and complete a baseline assessment followed by monthly surveys.

"At Apple, we believe responsible innovation starts with scientific rigor and clinical grounding. Health is deeply personal, and 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's vice president of Health and Fitness, said in a statement. Desai added that the findings will help the company build trusted tools that support individuals in managing their well-being.

The study also seeks to close the communication gap between patients who search for symptoms at home and the doctors who treat them in clinics.

"Patients and clinicians are both turning to AI, but the tools they use are being built and tested in separate worlds," said Robert M. Wachter, M.D., professor and chair of the Department of Medicine at the University of California, San Francisco. "When patients and doctors draw on versions of the same trustworthy tools, patients arrive prepared, conversations improve, and next steps happen seamlessly. Connecting these tools, and rigorously studying what works, is how we'll get there."

Personal Biometrics and Privacy Safeguards

A central component of the research involves testing how personal health metrics influence the relevance and accuracy of AI-generated guidance. Participants have the option to share specific data categories from Apple's HealthKit hub—such as resting heart rate, daily physical activity, and sleep patterns collected via iPhone or Apple Watch—to receive more tailored responses.

Furthermore, if a participant has imported clinical documentation into their device's Health app and grants explicit permission, the study app can incorporate those records, primarily laboratory test results, when answering questions.

Because consumer health technologies powered by large language models can occasionally produce incomplete or unclear answers, Apple stated that the app is designed to guide users toward qualified healthcare providers when queries involve concerning medical issues. Researchers will closely monitor how the chatbot complements the traditional healthcare system and facilitates the handoff to human physicians.

To address privacy concerns surrounding sensitive medical records, Apple and OpenEvidence have implemented strict data-handling protocols. According to TechCrunch, all user data is anonymized before being transmitted to a secure third-party cloud provider, and OpenEvidence deletes all query and response data after 24 hours.

Both CNET and TechCrunch report that information gathered during the study will not be used to train any artificial intelligence models, though the insights gained may inform Apple's future product and feature development.

Eligibility Requirements and How to Participate

The Apple Health AI Usage Study is currently open for enrollment in the United States, with the dedicated app available for download now. While Fierce Healthcare and CNET note that participants may remain in the study for up to 12 months, TechCrunch reports the company aims to run the study for more than 12 months.

To qualify for one of the 200,000 participant slots, individuals must meet specific criteria:

  • Be at least 21 years of age and reside in the United States.
  • Use an Apple Intelligence-enabled iPhone running iOS 26 or later. Regarding specific hardware, CNET and Fierce Healthcare specify that an iPhone 15 Pro or later is required, while TechCrunch describes the requirement as an iPhone 15 or later with Apple Intelligence enabled.

According to Apple, the advanced hardware requirement is necessary because the study app relies on the local processing capabilities of on-device Apple Foundation Models. These built-in models perform tasks directly on the handset, such as classifying the categories of health questions participants ask, before routing clinical queries to the OpenEvidence platform.

Sources