Incident sheet · SIG-8070 · /enterprise-ai

Signal
SIG-8070
Filed
Read
3 min · 568 words
Severity
Routine

Apple Bets on a Study, Not a Product: 200,000 Users to Test Clinical AI

Apple partnered with OpenEvidence on a health AI study app for up to 200,000 U.S. users over 12 months, with clinically grounded answers and strict data limits — but no product commitment.

Cam feed · SIG-8070ENT
Apple launches new app to study how people use AI for health
FeedApple launches new app to study how people use AI for health — Jonas Leupe / unsplash

Event log

  1. Apple launched a health AI study app today with OpenEvidence, targeting up to 200,000 U.S. participants for over 12 months.
  2. Eligibility requires age 21+, iPhone 15 or later, iOS 26 or later, and Apple Intelligence enabled.
  3. OpenEvidence deletes query and response data after 24 hours; Apple says data is anonymized and not used for AI training.
  4. OpenAI reported over 300 million health-related queries per week earlier this year.

Apple wants up to 200,000 U.S. participants to test health AI for more than 12 months. The company announced a new study app, built with healthcare-focused AI firm OpenEvidence, and released it for download today. What Apple did not announce is a product.

The study targets iPhone users over 21, running iPhone 15 or later hardware, iOS 26 or later, with Apple Intelligence enabled. That hardware cutoff alone excludes a large share of the installed base. Participants will ask health-related questions, much as they already do with ChatGPT or Claude, but the answers will be grounded in clinical sources: The New England Journal of Medicine, the Journal of the American Medical Association, and Nature.

What does the app actually do?

The app delivers simplified answers to everyday medical questions, Apple said. Three mechanics stand out:

  • Users rate answers and give feedback on responses.
  • Participants complete monthly surveys.
  • Users can opt to share HealthKit data — resting heart rate, activity, sleep, lab results — for more personalized suggestions.

Apple claims the app can help users prepare for doctor visits and will route them to specific providers when questions touch on concerning medical issues. That routing claim carries no detail on how "concerning" is defined, who triages it, or what liability follows a missed escalation.

What happens to the data?

Apple specified that it anonymizes data before sending it to a secure third-party cloud provider, and that the data will not train any AI models. OpenEvidence deletes query and response data after 24 hours, the company said. A 24-hour deletion window still leaves operational questions: what counts as a query, whether logs persist elsewhere, and which third-party cloud provider holds the data. Apple did not name the provider.

Sumbul Desai, M.D., Apple's vice president of Health and Fitness, framed the move as research discipline rather than a market entry. "At Apple, we believe responsible innovation starts with scientific rigor and clinical grounding. 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," she said in a statement. She added: "Insights from this study will help advance the broader scientific conversation and ensure that as technology evolves, we continue building trusted tools that genuinely support people wherever they are on their health journey."

That is two promises of rigor in one quote. The study design, peer review pathway, and publication plans went unmentioned.

Why study instead of ship?

OpenAI and Anthropic already field health-capable chatbots with specialized features. OpenAI said earlier this year that users submit over 300 million health-related queries every week — roughly 43 million a day, or 15.6 billion a year. That is the demand Apple is measuring before committing.

Apple has not said whether it intends to bring AI-powered answering to its Health app or to Siri. The company is, in effect, spending 12 months and a six-figure participant pool to decide what rivals already ship. The defensible read: Apple wants clinical grounding as a differentiator, and a study buys evidence, cover, and time.

The caveats remain. A 200,000-person, self-selected, iPhone 15-plus cohort will not represent the general population. Twelve months is a long watch window in a market moving at weekly release cadence. And no anonymization guarantee has stopped a breach yet.

via techcrunch.com (Source)

Filed under

  • apple
  • clinical-ai
  • health-ai
  • openevidence
  • data-privacy
Share this article:

More from Daniel Okafor

Daniel Okafor

Show full bio

Staff writer covering technology and product news in enterprise information technology and the leaders who run at CIO Monitor.

148 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous article