Apple (AAPL) Recruits Up to 200,000 Americans for Health AI Study
Apple (AAPL) and OpenEvidence will recruit up to 200,000 Americans to rate medical AI chatbot answers, with citations from JAMA, Nature and other journals.
AI SummaryAI
- Apple and OpenEvidence launched a health AI study on 2026-11-15, recruiting up to 200,000 US participants.
- AI answers carry citations from the New England Journal of Medicine, JAMA and Nature.
- HealthKit sharing of heart rate, activity, sleep and lab results is optional and excluded from AI training.
- JAMA counted 952,008 US clinicians entering 271 million clinical questions on OpenEvidence between January 2024 and June 2026.
Study Seeks Up to 200,000 US Participants
Apple (AAPL) has launched a United States research programme that will recruit up to 200,000 people to rate how trustworthy and how understandable answers from medical artificial intelligence chatbots are, the company said in its announcement on 2026-11-15. The programme, called the Apple Health AI Usage Study, runs entirely online across the country and enrolls participants through a dedicated app. In the app, a participant types an everyday health question, reads the answer the AI generates, and rates it. The study also tests whether complex medical information is understood when it is delivered in plain language.
Apple listed four focus areas in the announcement: user trust in the answers, answer quality, comprehension, and effects on everyday health behaviour. One further question asks whether medical AI helps users prepare their questions ahead of a consultation. Every answer carries citation information, so a participant can check the underlying medical literature. OpenEvidence, the study partner, generates its answers from sources including the New England Journal of Medicine, JAMA and Nature. The scope reaches into real appointments as well. Apple said it will examine how participants use AI to understand health information and organise their questions before they see a doctor.
Participation is voluntary at each step. A participant may share HealthKit data such as resting heart rate, activity levels, sleep and laboratory results, but sharing is optional, and answering survey questions is a separate decision. Apple said data provided through the study will not be used to train AI models. The app also carries a safeguard that detects concerning questions and directs users to contact a medical professional or another source of help. The company states that the app is not a diagnostic or treatment tool and does not replace advice from clinicians. Users dealing with a health problem, or checking medical information, are told to consult a doctor or clinical specialist.
271 Million Clinical Questions Logged
The design marks a departure from Apple's earlier health research. In February 2025 the company began the Apple Health Study with Brigham and Women's Hospital, drawing on iPhone and Apple Watch data to examine links between physical and mental health. That earlier work centred on the signals devices themselves capture, such as movement and heart data; the 2026 programme centres on language, the medium through which most people now first meet medical information. Both programmes sit in the same research apparatus, but they sample different evidence.
For OpenEvidence, the study extends an already active clinical footprint. The company earlier announced a partnership with Anthropic to provide its clinical decision-support AI free of charge to medical professionals in roughly 100 low- and middle-income countries. Usage on the clinician side is already large enough to benchmark, and the numbers are public. A JAMA paper published on 2026-10-15 analysed activity on the platform and counted 952,008 United States clinicians entering 271 million clinical questions between January 2024 and June 2026. Of those questions, 51.9% related to clinical decision-making and 33.1% concerned treatment or management.
Those figures come from the platform's own usage records. They establish how deeply the tool has entered clinical work in the United States, but they measure neither how a patient without medical training judges an answer nor any effect on medical performance. The JAMA analysis flags that limit itself, and no comparable public dataset covers the consumer side. Whether an answer written for a clinician's vocabulary reads the same way to someone managing a condition at home remains, for now, untested. Apple's study supplies that measurement directly, on a participant pool of up to 200,000 people rating trust and comprehension.
AAPL Trades Under Strong Resistance
COINOTAG data shows Apple (AAPL) trading at $357.53 on Monday, up 0.39% over 24 hours within a $356.13 to $361.12 range, with the daily trend up and the RSI reading 61.4. The composite scoring rates the nearest resistance, $361.23, at 86 out of 100; the level coincides with the upper Bollinger Bands boundary. The strongest support below the market is $349.69, also scored 86. The stock therefore sits beneath a strongly scored resistance a little above the current price, while $349.69 marks the first floor a pullback would meet; a close above $361.23 would put the moderate resistance at $364.95 in view.
AI-generated, AI-reviewed, under COINOTAG editorial oversight.

