Apple announced on Sunday that it is partnering with clinical‑AI platform OpenEvidence to launch the Health AI Usage Study, a research effort aimed at understanding how Americans trust and use AI‑driven health chatbots on their iPhones. The study will collect data on user behavior, comprehension of AI responses, and perceived answer quality, with the goal of informing responsible health‑AI design for Apple and the broader industry.

Study Design and Participation Criteria

According to the HLTH conference briefing, the study will accept up to 200,000 participants who are at least 21 years old, own an iPhone 15 Pro or newer, and run iOS 26 with Apple Intelligence enabled. Participants will access the study through an app built by OpenEvidence, a tool already used by clinicians for evidence‑based AI assistance. The research will span more than a year, featuring a monthly poll that asks users about the steps they take for their health, how well they understand AI answers, and how useful those answers are.

Data Collection, Privacy Measures

Apple emphasizes that any HealthKit data shared—such as resting heart rate, sleep, activity, or lab results—will remain outside model training pipelines. Queries are anonymized before being sent to a secure third‑party cloud, and OpenEvidence deletes both questions and answers after 24 hours. Participants may also opt to share survey responses, but all data will be used solely for research and not for improving Apple’s AI models.

Context and Industry Implications

The initiative comes as consumer health chatbots see rising usage; OpenAI reported that users ask more than 300 million health‑related questions weekly. Apple’s move seeks to address concerns that large language model (LLM) health tools can provide incomplete or unclear information, a point highlighted by recent lawsuits alleging harmful advice from AI chatbots. By linking each AI response to peer‑reviewed citations from journals such as Nature, NEJM, and JAMA, the study aims to set a benchmark for transparency and reliability in health AI.

What to Watch Next

The first batch of monthly survey results is expected later this year, which could reveal patterns in user trust and the impact of personal health data on answer relevance. Industry observers will be watching whether Apple’s findings influence broader regulatory guidance for AI‑driven medical tools and whether other device makers adopt similar evidence‑based approaches.

FAQ

Who can join Apple’s Health AI Usage Study?

U.S. residents aged 21 or older who own an iPhone 15 Pro or newer with Apple Intelligence and iOS 26 or later can sign up.

Will the data collected be used to train Apple’s AI models?

No. Apple states that all user queries are anonymized, deleted after 24 hours, and excluded from any model‑training processes.

What kind of health information can participants share?

Participants may optionally share HealthKit metrics such as resting heart rate, sleep, activity, and lab results, as well as survey responses about their health‑AI interactions.