Apple opens a study of up to 200,000 people on how they use AI for health questions
The Apple Health AI Usage Study, built with clinical AI platform OpenEvidence, asks US adults to put medical questions to an AI and rate the answers for accuracy, detail and relevance.
Apple has launched the Apple Health AI Usage Study, a virtual research study with its own iOS app. It is built with OpenEvidence, a clinical AI platform used by healthcare professionals. Participants ask the app health questions, read the answers and rate them. Apple is looking for up to 200,000 adults in the US. The app is available now and enrollment opened on November 15.
Many people already ask chatbots about symptoms, lab results and medication. Little is known about how they read those answers or what they do next. Apple is trying to measure that at scale before it builds anything further. In its announcement, Apple says it wants to understand what makes people trust these tools and whether the answers are good.
What the study measures
Apple says the study looks at four things: how much users trust the tool, how good the AI responses are, how well people understand health information, and what everyday health actions they take. Participants rate responses in surveys. The Verge reports that the ratings cover accuracy, detail and relevance.
The study also has two narrower questions. The first is whether answers become more relevant and easier to understand when people connect their own health data. The second is whether AI can help people prepare for a doctor's appointment. The second question puts the tool in the time before a visit, as preparation for the doctor rather than a replacement.
Apple says the app does not diagnose or treat conditions and does not replace medical advice. It also says the app has guardrails that detect certain concerning questions and suggest that the user contact a provider or other resources.
How the app works
OpenEvidence provides the AI technology that generates the answers, and Apple designed the research experience. According to Apple, OpenEvidence draws on established medical sources, including The New England Journal of Medicine, the Journal of the American Medical Association and Nature. Each response comes with linked citations to peer-reviewed sources.
The app takes text prompts. The Verge reports that dictation works through the iOS keyboard, but the app is not connected to Siri, and the AI is not Apple Intelligence. Participants need to be 21 or older, live in the US, and have an iPhone 15 Pro or later running iOS 26 or later. Apple describes these as Apple Intelligence-enabled iPhones.
During enrollment, participants can choose to share HealthKit data, such as resting heart rate, activity, sleep or clinical lab results, so that answers can take their own health into account. Sharing survey responses is also optional. Apple says study data will not be used to train AI models, and that it has security measures to protect the data it holds and participants' confidentiality.
The announcement does not say how long the study will run, how or when results will be published, or whether it will lead to a feature in Apple's Health app.
Why Apple frames it as research
"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," said Sumbul Desai, Apple's vice president of Health and Fitness. She said the insights would help "advance the broader scientific conversation."
Apple also quotes Robert M. Wachter, chair of the Department of Medicine at UCSF. He says patients and clinicians are using AI tools that "are being built and tested in separate worlds." In his view, connecting those tools and studying them rigorously is the way forward. The study fits that description. A tool that clinicians already use is being given to the public, with patient data linked to it, and people's behaviour is measured rather than assumed.
The study's design also has limits. The ratings come from participants, not from clinicians checking the answers. A person can rate an answer as clear and useful when it is wrong for their situation. The participants will also lean toward people who own a recent iPhone and choose to join a health study, so they will not represent all patients.
What it means for work
A large share of health information work currently falls on clinicians, nurses and pharmacists: explaining lab results, answering follow-up questions and helping patients prepare for appointments. Tools like this one are designed to take over the first round of that explaining before the patient meets a professional. If the study shows that people arrive at appointments better prepared, clinicians may spend less of a visit on basic explanation and more on decisions. If it shows that people trust confident answers more than they should, clinicians will have more work correcting what patients believe. For the people who build these tools, the study changes how quality is tested, from benchmark scores to how people actually understand and act on an answer. How much anyone can rely on the answers is still open until Apple publishes results, and the announcement gives no date for that.
Sources
- Apple's new study looks to improve the effectiveness of AI health toolsThe Verge
- Apple launches research study to explore the use of AI for healthApple Newsroom
This article was researched, written and illustrated by AI, and published without human editing. Spotted an error? Let us know.
More from māchine
All storiesGoogle has added computer use to the Gemini app for macOS. Paying US subscribers can hand it multi-step tasks, decide which apps it may touch and approve its key actions.
Yale researcher Julián Posada argues that platforms recruiting doctors, lawyers and PhDs to train AI pay well at first, then cut rates and drop workers, as they did in Venezuela and Kenya.
Microsoft offered voluntary exits to about a quarter of its workforce in October. Its CEO says the reason is AI, as a reorganisation of work around managing agents, not as a replacement for people.