Apple launched a new research initiative Sunday aimed at evaluating how everyday users interact with conversational artificial intelligence health tools and whether the responses provided are useful, accurate and easy to comprehend.
The tech giant is recruiting up to 200,000 participants in the United States for the Apple Health AI Usage Study, which operates through a dedicated application now available in the App Store, according to reports from CNET and ABC News. To join, participants must be at least 21 years old and possess an iPhone 15 Pro or later model running iOS 26 or higher.
How the AI Study Works
Through the study app, users can submit text-based questions regarding general health, medical symptoms, or recovery from surgical procedures. The software returns responses written in accessible language, accompanied by direct links to peer-reviewed clinical literature, CNET and Engadget reported. To generate these answers, Apple is collaborating with OpenEvidence, a clinical AI platform used by medical professionals that draws from major publications including The New England Journal of Medicine, the Journal of the American Medical Association, and Nature.
Participants can also choose whether to share personal HealthKit information—such as sleeping habits, physical activity levels, blood test results, and resting heart rates—to receive tailored guidance, according to ABC News. For instance, a user seeking to lower their resting heart rate could obtain tailored recommendations incorporating their own recorded metrics rather than generic suggestions.
Dr. Sumbul Desai, vice president of Health and Fitness at Apple, told ABC News that the goal is to examine how AI can bridge patient care. “When people have access to clear, medically grounded information, they can walk into their doctor’s appointments feeling more prepared, informed, and confident,” Desai said, emphasizing the need for scientific rigor and clinical grounding.
Privacy Controls and Safety Guardrails
Privacy measures are integrated into the processing flow, according to company details reported by CNET. When a user submits a question, the iPhone uses tools to strip away personally identifiable details. The prompt is then transmitted to a cloud environment operated by third-party contractor UST Global, where queries are unlinked from participant IDs before being sent to OpenEvidence. OpenEvidence deletes all question and response records within 24 hours and will not use participant data to train AI models.
Apple also emphasized that the study application includes safety features designed to identify concerning queries and suggest that users contact healthcare professionals, according to ABC News. However, the app does not feature live medical monitoring, and research staff will not contact users directly if questions related to self-harm are submitted, CNET reported.
Addressing Gaps in Digital Health
The initiative comes as public interest in AI health assistants grows, alongside concerns over inaccurate guidance. A study published in Nature Medicine in February 2026 found that while AI chatbots correctly detected medical conditions nearly 95 percent of the time during clinical testing, regular users arrived at the correct conclusion less than 35 percent of the time when attempting to understand dense guidance, ABC News reported. Meanwhile, Pew Research Center data indicates that 35 percent of U.S. adults already use AI chatbots for health information.
Dr. John Brownstein, a professor of pediatrics and biomedical informatics at Harvard Medical School and an ABC News contributor, noted that converting personal health metrics into actionable understanding remains a major hurdle. “We need to understand not just whether AI provides accurate answers, but whether people understand those answers, trust them appropriately and ultimately make better health decisions,” Brownstein told ABC News.
Enrolled participants will be asked to remain in the study for up to 12 months, completing monthly surveys and rating the answers they receive, CNET reported. Apple has not announced a publication date for the study’s findings, but the company stated it plans to share the research broadly to help advance digital health communication.



