OpenAI is flipping the switch on ChatGPT Health for everyone in the US this Thursday, letting far more people plug their medical records and health-tracking data straight into the chatbot. According to The Verge AI, the company paired the rollout with a striking claim: OpenAI’s vice president of health product, Ashley Alexander, said its models “are now capable of reasoning at levels that are better than clinician level.” That’s a big statement, and OpenAI’s own team walked it back within the same briefing.
When pressed on how the models actually stack up against human clinicians, OpenAI health lead Karan Singhal said he’d “temper” the claim, noting only that “there have been individual studies that have been pointing in that direction,” like one from Harvard and Stanford. So the headline number and the fine print don’t quite line up. That gap matters when the product touches people’s health decisions.
What’s actually changing
- You can ask about your connected medical info directly in the regular chat window.
- ChatGPT asks permission before using your health data to personalize a response.
- OpenAI says “all conversations are encrypted at rest and at transit,” with health data getting “additional encryption protection.”
- You can review and change what you share: post-visit notes, care team notes, lab results, and medications.
Data sources include your health providers plus Apple Health, Weight Watchers, MyFitnessPal, and Function. It’s rolling out to logged-in users 18 and older on web and iOS, across the free, Go, Plus, and Pro plans. It runs on OpenAI’s newest models, and The Verge AI notes the release lands just weeks after the company launched GPT-5.6 Sol, which it billed as its “strongest model yet for health.”
Why the timing is loaded
Here’s what stands out. One day before this launch, a Florida pastor sued OpenAI, alleging ChatGPT gave him “extremely dangerous medical recommendations” that led him to delay care for a pulmonary embolism. A pulmonary embolism is a blocked artery in the lungs, and it can be fatal without prompt treatment. Shipping a broad health expansion the day after that suit surfaces is a bold move, and it puts OpenAI’s “better than clinician level” framing under a harsher light.
OpenAI is careful to say ChatGPT Health “supports, not replaces, professional care.” That’s the guardrail the whole product leans on. But the tension is obvious: market the reasoning as clinician-grade, then remind everyone it’s not a doctor.
Why this matters
This is significant because it moves consumer AI from general questions into personalized medical territory at scale. Connecting real lab results and medication lists is a different risk category than asking ChatGPT to summarize an article. When the model reasons over your actual records, a confident wrong answer carries real weight.
For practitioners and builders, a few things are worth watching:
- The claims gap. Executives are making clinician-level assertions that the research leads themselves soften. Expect regulators and clinicians to push on the evidence.
- Liability is now live. The Florida lawsuit is an early test of who’s responsible when an AI health answer goes wrong. The outcome could shape how every health-AI product is built and disclosed.
- Data trust. Encryption promises are table stakes now. The bigger question is how OpenAI uses connected health data over time, and how clearly it explains that to users.
What to expect next: closer scrutiny of the studies OpenAI cites, more explicit disclaimers baked into the chat flow, and likely a wave of competitors matching the feature. If you’re using ChatGPT Health, treat it as a second opinion that helps you ask better questions, not a substitute for the person who can actually order the test. More detail is available in the original report from The Verge AI.