Apple launches study to test trust in AI health answers

Apple and OpenEvidence launched a US study with up to 200,000 participants to rate AI health answers via an app. The study focuses on trust, comprehension, and privacy. The app is not a diagnostic tool.

Apple has launched the Apple Health AI Usage Study in partnership with OpenEvidence, aiming to enroll up to 200,000 US participants aged 21 and older. The study is accessible through a dedicated iOS app, where users can ask health-related questions and rate AI-generated responses for accuracy, detail, and relevance. OpenEvidence grounds its answers in peer-reviewed medical literature, drawing from reputable sources such as Nature, The New England Journal of Medicine, and JAMA. The research focuses on user trust, comprehension, response quality, and the impact of AI on health decision-making. Participants may optionally share HealthKit data to determine if personal context improves answer relevance. Apple emphasizes that the app is not a diagnostic tool and does not replace medical professionals. Privacy protections are in place, and study data will not be used to train AI models. The project is expected to last over a year, with monthly surveys to collect ongoing feedback.