Apple has opened a quiet research door in the United States, asking a simple and very old human question in a new form: when ordinary people carry their worries about the body to a machine, does the answer help them understand and steward their own health? Working with OpenEvidence, a health AI company, Apple has launched a study app that will run for more than a year and seek up to 200,000 participants, each of whom will test answers grounded in medical literature and report back on whether those answers are worth trusting. The experiment arrives as the chatbot has already become a common fir
Apple launches Health AI Study app with OpenEvidence to test how people use AI for medical questions
Apple wants to know what happens when people bring health worries to an AI.
Apple is running a study instead of just shipping a health chatbot. Why does that matter?
It tells you Apple wants evidence on how people actually use AI for medical questions first. Up to 200,000 participants, over more than 12 months, is a large sample.
And the answers come from OpenEvidence, not a general chatbot?
Right. The answers are grounded in clinical literature, with sources like NEJM, JAMA and Nature named.
Worth noting that's Apple's description of the design. The reporting doesn't show any independent check of how accurate the answers are.
What about the personal data?
Participants can choose to share things like resting heart rate, sleep, activity and lab results through HealthKit. Apple says that makes answers more individual and can help with doctor visits.
The privacy claims are also Apple's: de-identified data, no AI training, and OpenEvidence erasing queries within 24 hours. They're stated commitments, not verified ones.
Does this mean a Siri health feature is coming?
Apple hasn't said so.
Right, so any talk of a Siri or Health app tool is speculation. The only context we have is OpenAI's claim of over 300 million weekly health questions, which is OpenAI's own figure.
Le Pouls
- Millions of people already type their fears about symptoms into chatbots, and Apple is now trying to measure whether that habit can be made safer and more useful.
- Eligibility is tightly drawn to adults 21 and older with iPhone 15 or later, iOS 26 and Apple Intelligence, so the study is starting narrow before it widens its reach.
- Participants will rate AI answers, give feedback and complete monthly surveys, turning everyday health curiosity into a steady stream of evidence.
- Privacy guardrails, including de-identified data, a promise that health data will not train AI models, and a 24-hour erasure window at OpenEvidence, are being offered to calm concerns about sensitive information.
- The outcome is still open: Apple has not said whether findings will shape an AI feature in its Health app or Siri, leaving the next move unknown.
Apple has opened a quiet research door in the United States, asking a simple and very old human question in a new form: when ordinary people carry their worries about the body to a machine, does the answer help them understand and steward their own health? Working with OpenEvidence, a health AI company, Apple has launched a study app that will run for more than a year and seek up to 200,000 participants, each of whom will test answers grounded in medical literature and report back on whether those answers are worth trusting. The experiment arrives as the chatbot has already become a common first stop for medical uncertainty, and it leaves unresolved what Apple will ultimately build from what it learns.
Apple has released a research app in the United States that asks what happens when people bring everyday health worries to an AI. The Health AI Study app launched on Sunday and is open for now only to qualified US participants. Built with OpenEvidence, it aims to learn whether AI answers grounded in clinical evidence can help people understand health information and manage their own care, over more than twelve months and up to 200,000 people.
To join, participants must be at least 21, own an iPhone 15 or later, run iOS 26 or later, and have Apple Intelligence turned on. Inside the app they ask ordinary health questions, judge the answers, and fill in monthly surveys about how useful they find AI medical guidance. Answers draw on sources such as The New England Journal of Medicine, JAMA and Nature, and are simplified so that health topics are easier to grasp. Those who choose can share data from Apple's HealthKit, such as heart rate, sleep and lab results, to make responses more personal and help prepare for doctor visits. When a question points to something concerning, the app directs users toward a healthcare provider.
Apple stresses privacy: data is de-identified, will not be used to train AI models, and OpenEvidence says it erases query and response data within 24 hours. The timing matters, since OpenAI and Anthropic have already built health features into ChatGPT and Claude, and OpenAI recently reported more than 300 million health questions a week. What Apple will do with the findings remains unclear, and whether they lead to a tool in the Health app or Siri is the question to watch.
Apple has released a research app in the United States built to answer a plain question: what happens when people bring their everyday health worries to an AI? The Health AI Study app was launched on Sunday, with a download date of 16 November on Apple's US App Store, and it is open only to qualified US participants for now.
The company built it with OpenEvidence, a health AI company. Together they want to find out whether AI answers backed by clinical evidence can help people understand health care information and manage their own health. The study is set to run for more than 12 months and aims to enrol up to 200,000 participants.
Eligibility is narrow. Participants must be at least 21, own an iPhone 15 or later, run iOS 26 or later, and have Apple Intelligence switched on. Inside the app, they ask ordinary health questions, evaluate the AI-generated answers, rate them and give feedback. They also take monthly surveys meant to gauge how valuable they find AI-provided medical guidance. Apple says the results will help it understand how people interact with AI-driven health information.
The app is meant to differ from a general-purpose chatbot. Its answers are grounded in medical research and clinical literature, with sources that include The New England Journal of Medicine, the Journal of the American Medical Association (JAMA) and Nature. Apple says the design aims to simplify medical information so health topics are easier to grasp.
Participants can also choose to share health and wellness data through Apple's HealthKit platform. That could include resting heart rate, physical activity, sleep records and laboratory results. Apple says this can make responses more individual and help people prepare for medical visits. When a question touches on a potentially concerning condition, the app can point the user toward healthcare providers.
Privacy is a central part of the pitch. Apple says participant information is de-identified before it goes to a secure third-party cloud data service provider, and that it will not be used to train AI models. OpenEvidence, according to Apple, will erase users' query and response data within 24 hours. Sumbul Desai, MD, Apple's vice president of Health and Fitness, said in a statement that responsible innovation at Apple starts with scientific rigor and clinical grounding.
The launch lands as other AI companies move deeper into health. OpenAI and Anthropic have both introduced dedicated health experiences through ChatGPT and Claude. Earlier this week, OpenAI said its users were asking more than 300 million health-related questions every week, a figure that shows how much medical information-seeking has already moved to chatbots.
What Apple does with the findings is unanswered. The company has not said whether the study will lead to an AI-powered tool in its Health app or in Siri, and that is the thing to watch as the study runs.
Citations marquantes
Responsible innovation starts with scientific rigor and clinical grounding.— Sumbul Desai, MD, Apple vice president of Health and Fitness (paraphrased)