ChatGPT Health Is Now Open to All U.S. Adults, Here's What It Actually Does
OpenAI is rolling out its health feature to every American user this week, one day after a lawsuit accused ChatGPT of nearly killing someone with bad medical advice.
Key points
- OpenAI opened ChatGPT Health to all U.S. Users aged 18 and over on free, Go, Plus, and Pro plans as of this week.
- Users asked 300 million health-related questions per week on ChatGPT as of May 2025, up from 230 million in January 2025.
- A Florida pastor filed a lawsuit against OpenAI one day before the launch, alleging ChatGPT told him not to see a doctor for a condition that later proved nearly fatal.
- OpenAI's own terms still say the service is "not intended for use in the diagnosis or treatment of any health condition."
- GPT 5.6-Luna, OpenAI's smallest new model, outperforms GPT 5.5 on HealthBench, an open benchmark for testing AI on medical questions.
OpenAI launched ChatGPT Health to every logged-in American adult this week, first reported by TechCrunch AI. The feature lets you connect personal health data and ask ChatGPT questions about it, inside the same chat window you already use. No separate app, no waiting list.
The timing is hard to miss. One day before the rollout, a Florida pastor sued OpenAI, claiming ChatGPT told him not to bother seeing a doctor for a symptom that turned out to be serious. OpenAI cited its own terms of service in response. Those terms state plainly that ChatGPT is "not intended for use in the diagnosis or treatment of any health condition." The company told the New York Times it's working on making medical answers safer. Given what we found when we reported on health data privacy on 15 July, the question of who controls your connected records here deserves more than a checkbox in settings.
What does ChatGPT Health actually connect to?
You can link Apple Health and MyFitnessPal, plus the health tracking service Function. Medical records come in from hospital systems like Epic and Oracle Health, and from services including One Medical and Function Health.
Until now, health questions using that data had to go through a dedicated section inside the app. OpenAI found that 70% of health-related questions happened in regular chat instead, so it changed the setup. Your health data now flows into any conversation where it's relevant, whether you're asking about food or a test result.
OpenAI says it doesn't use your health data to train its models, and that it works with physicians to improve how its AI handles medical questions.
Should you trust it with real medical decisions?
No, and OpenAI says so itself. The terms are unambiguous: this isn't a diagnostic tool.
That caveat matters because 300 million health questions a week tells you people are already using it this way. Multiple independent studies have found that AI chatbots give unreliable medical advice often enough to cause harm. Competitors including Anthropic and Google have launched similar features despite those findings.
| Feature | Detail |
|---|---|
| Rollout | U.S. Users, web and iOS, this week |
| Plans covered | Free, Go, Plus, Pro |
| Data sources | Apple Health, MyFitnessPal, Function, Epic, Oracle Health, One Medical |
| Weekly health queries | 300 million (May 2025) |
| Benchmark model | GPT 5.6-Luna, beats GPT 5.5 on HealthBench |
Common questions
Is my health data safe with OpenAI?
OpenAI says it doesn't train its AI models on your personal health data. You control which apps and records you connect, and you can disconnect them at any time in the app's settings.
Can ChatGPT replace my doctor?
No. OpenAI's own terms state the tool isn't designed to diagnose or treat any condition. Use it to understand information or prepare questions for a real appointment, not to decide on treatment.
Who can use ChatGPT Health right now?
Any U.S. Resident aged 18 or older with a free or paid ChatGPT account, on the web or iPhone, can access the feature this week.
Connect your health data if it helps you understand your own information. But anything ChatGPT tells you about symptoms or treatment belongs in front of a doctor before you act on it. The 300 million weekly questions are the real story here: the product is already being used as a medical tool, whatever the terms say.
