Apple launches AI health study with OpenEvidence

Apple's Health AI Usage Study lets up to 200,000 US iPhone users ask health questions answered by OpenEvidence, to test trust and use.
What it means for founders
Apple is not launching a consumer health product. It is running a research study first, while OpenAI and Anthropic already offer health features. TechCrunch notes OpenAI said earlier this year that users ask more than 300 million health questions a week.
If you build consumer health AI, this study could become the reference for what "trustworthy" means in practice: cited answers, opt-in personal data, short data retention and guardrails that route people to doctors. Design to that bar now rather than retrofit later. Microsoft has also tested AI on hard diagnoses; expect evidence to become the price of entry.
For OpenEvidence, the collaboration moves a clinician tool toward consumers through Apple, a path other clinical AI startups will want to copy.
Watch for whether Apple publishes results and whether health answers later show up in Siri or the Health app.
What we don't know yet
Apple has not said when results will be published or whether the study will lead to a product. TechCrunch notes it is not clear if Apple plans AI answers for its Health app or Siri. Neither Apple nor OpenEvidence has said whether the collaboration involves payment.
The story
Apple has launched the Apple Health AI Usage Study, a US research study and app that answers everyday health questions using OpenEvidence, a clinical AI platform widely used by healthcare professionals, the company announced on November 15. Apple says the aim is to understand trust, response quality, comprehension and everyday health actions when people use conversational health AI.
Up to 200,000 US adults aged 21 or older can join. They need an iPhone 15 Pro or newer on iOS 26 or later, with Apple Intelligence turned on.
How the app works
Participants ask health questions and get answers in plain language with linked, peer-reviewed citations. OpenEvidence builds its answers from medical literature including JAMA, Nature and The New England Journal of Medicine, Apple says. People can opt to share HealthKit data such as lab results, sleep, activity or resting heart rate to see whether personal data makes answers more relevant. The study will also explore whether health AI can help people prepare for visits with their own doctors. Apple says safety guardrails are intended to detect certain concerning questions and suggest following up with a provider or other resources. The app is not meant to diagnose or treat any condition, Apple says.
TechCrunch reports that participants can rate answers and will be asked to take monthly surveys, and that Apple aims to run the study for more than 12 months.
On privacy, Apple says nothing from the study will train AI models. TechCrunch reports that Apple said data is stripped of identity before it reaches an outside cloud provider, and that OpenEvidence deletes both questions and answers once 24 hours have passed.
"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," said Sumbul Desai, M.D., Apple's vice president of Health and Fitness.
Sources
Primary sources
Reporting
More in Products & Launches
- Amazon memo maps a product graph to counter AI agentsReuters says Amazon's World Product Graph would cover virtually every item for sale on Earth. Amazon calls the memo an outdated brainstorm.The Verge · 4d ago

- llama.cpp hits v1.0.0 with a versioned C APIllama.cpp and ggml reach v1.0.0, tying breaking C API changes to major version bumps after pruning 35 deprecated functions.Hugging Face · 5d ago

- ElevenLabs launches ElevenAI note-taker for teams

For founders: Teams picking a note-taker: the free window makes a pilot cheap, but ElevenLabs lists no price after it ends.
TechCrunch · 5d ago - OpenAI launches ChatGPT Wallet for agent purchasesChatGPT Wallet lets OpenAI's dots agents buy for you with tokenized cards and Face ID or fingerprint approval.OpenAI · 5d ago

- Google launches Gemini Decisions API

For founders: Pack your checks into one call. Google bills the shared input once per request, however many questions ride on it (each adds only its own instruction tokens)…
Google AI · 7d ago