Artificial intelligence (AI) chatbots are increasingly becoming a go-to source for people seeking answers to their health-related questions. From understanding symptoms and interpreting blood test results to improving sleep and managing fitness, people are turning to AI for quick medical information. However, concerns about inaccurate advice, misleading responses and potential health risks have also raised questions about how reliable these tools really are. In a move to explore how AI can make health information more trustworthy and accessible, Apple has launched the Apple Health AI Usage Study app in the US. The research initiative aims to understand how people use AI for health-related queries and whether medically grounded, personalised responses can help them make better-informed health decisions. According to Apple, the study app will answer everyday health questions using trusted medical research and, with users' permission, information from their existing health records and metrics. It will also provide links to peer-reviewed clinical research, allowing users to understand where the information comes from. Now, access to clear, medically grounded information may help people feel better prepared and more confident when consulting their doctors. The study will examine whether AI can act as a bridge between patients and healthcare professionals rather than becoming a barrier to appropriate medical care.
How Will Apple's AI Health Study App Work?
Participants will be able to ask the app everyday health questions and receive responses supported by linked clinical citations. The idea is to make medical information easier to understand while allowing users to verify the research behind the answers. Apple is collaborating with OpenEvidence, an AI-powered medical information platform used by healthcare professionals. The platform draws on medical journals and other scientific literature to provide research-backed information. While OpenEvidence primarily serves healthcare professionals and delivers information in a format suited to clinical practice, Apple says it is using its own technology to translate medical research into simpler, more accessible responses for everyday users.
READ | Delhi's Toxic Air: Can Pollution Affect Your Appetite And Digestion Too?
The study will also explore how personal health information can make AI-generated responses more relevant. Participants can choose whether to share information already available through Apple's Health app, including resting heart rate, physical activity, sleep patterns and blood test results. For instance, someone asking how to improve their resting heart rate could receive suggestions informed by their own health measurements rather than relying entirely on general recommendations.
Why Is Apple Studying AI-Based Health Advice?
The growing popularity of AI chatbots for health information has highlighted an important challenge: providing an accurate answer does not necessarily mean users will understand it or know what to do next.
A February 2026 study from the University of Oxford, published in Nature Medicine, reportedly found that AI chatbots could identify medical conditions with nearly 95 per cent accuracy when tested on clinical cases. However, everyday users reached the correct answer less than 35 per cent of the time when attempting to interpret complex medical guidance. The gap highlights why medical accuracy alone may not be enough when developing AI tools for public healthcare use.
According to Pew Research Center data cited in the report, 35 per cent of adults, including around half of adults under 30, are already using AI chatbots. The study will therefore assess more than the accuracy of AI-generated responses. Researchers also want to understand whether people comprehend the information, trust it appropriately and use it to make better health decisions.
Can Apple's AI Health App Help Patients Prepare For Doctor Visits?
Another important objective is to understand whether AI-powered health information can improve communication between patients and their doctors. Participants will be asked to provide feedback on the responses they receive. They can rate the information and complete occasional surveys about their experience with the app. Researchers will examine whether these interactions help people understand their health measurements, identify relevant questions and prepare more effectively for clinical consultations. Dr Brownstein emphasised that collecting health information through smartphones and wearable devices is only one part of the challenge. Making that information understandable and actionable is equally important.
What About Privacy And Medical Safety?
Since the study may involve sensitive personal health information, privacy is an important part of Apple's approach. Apple says user questions will be processed through privacy protections designed to remove identifying information before they are shared with OpenEvidence. According to the company, OpenEvidence will delete question-and-response data within 24 hours and will not use these queries to train its AI model.
The app will also include safeguards designed to identify potentially dangerous situations and encourage users to seek professional medical assistance when necessary. However, Apple has clarified that the study app does not offer live medical monitoring and is not intended to replace doctors or other qualified healthcare professionals.
Who Can Participate In Apple's AI Health Study?
The study is currently open to adults aged 21 and above in the United States who own an Apple Intelligence-enabled iPhone, including the iPhone 15 Pro and later models, running iOS 26 or later. Eligible participants can download the dedicated study app, review the informed consent process and choose whether to enrol. Apple says the objective extends beyond developing future products. The company also intends to make the research broadly available to contribute to the future of AI-assisted health communication. While AI-powered health tools could make medical information more accessible, their usefulness will ultimately depend on whether people can understand the responses, recognise their limitations and seek professional care when needed.




