Apple today launched something a little different from its usual lineup of hardware and software updates. It’s called the Apple Health AI Usage Study, a research project built to understand how people actually behave when they turn to AI with a health question, and whether they end up trusting the answer.
The project runs through its own free app, available in the US, and it’s powered by OpenEvidence, a clinical AI platform that doctors already rely on. Instead of scraping random corners of the internet, it draws from peer-reviewed medical journals and recognized clinical research, so the answers are at least starting from a credible place.
What do you actually do inside the study?
Instead of searching for medical advice on the open web, participants can use the app and get answers with underlying research linked right there to check. They can also link HealthKit metrics like heart rate, sleep, or lab results to get even more personalized answers. Part of the goal is seeing whether the app leaves people a bit more prepared for their next real doctor’s visit, rather than just more worried.
Apple’s Sumbul Desai framed the reasoning simply, saying “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.”
That’s really the crux of why this study exists. AI chatbots still get health answers wrong sometimes, and when the subject is your own body, that’s not a small mistake to make.
Why should you care about the study?
Because whatever Apple learns here could shape how health AI behaves everywhere, including whatever eventually shows up on your iPhone. UCSF’s Dr. Robert Wachter put the stakes well, noting that when doctors and patients lean on the same trustworthy tools, “patients arrive prepared, conversations improve, and next steps happen seamlessly.”

Apple is also being upfront about privacy. Sharing HealthKit data is entirely optional, none of it trains outside AI models, and the app has built-in guardrails built to catch worrying questions and nudge you toward an actual doctor instead.
Several AI companies, including Microsoft with Copilot Health, Perplexity with Perplexity Health, and others, have launched similar products in the past. However, if there’s one company I would trust with my medical data, it would be Apple, and hopefully the study’s results will improve our understanding of our health without compromising our private data.

