ChatGPT Health now available: what changed overnight
ChatGPT Health is now available to the public after months of limited testing, letting users connect Apple Health and eligible medical records to view trends, browse data, and try suggested prompts. OpenAI says more than 300 million people already ask ChatGPT health questions weekly—making the wider rollout a major shift in how we seek care advice.
Key Takeaways
- OpenAI has opened ChatGPT Health to the public after an earlier tester-only phase.
- Users can connect Apple Health and eligible medical records to see trends, browse records, and try suggested prompts.
- More than 300 million people ask ChatGPT health questions weekly; about one in three U.S. adults used a chatbot for health info last year.
- A lawsuit filed the day before alleges an older ChatGPT model gave near-fatal advice; OpenAI says the tool is not for diagnosis or treatment.
- Clinicians urge careful prompting, fresh chats per concern, and pushback if answers flip too easily.
If you remember late-night WebMD spirals, printed lab sheets, and weeks waiting for a follow-up, this launch feels like a hard pivot. For years, people pieced together health answers from forums and appointment scraps. Now the chatbot millions already use can sit next to Apple Health data and eligible medical records. That is the then-and-now story behind why ChatGPT Health now available matters.
According to Mashable’s report on OpenAI’s Thursday announcement, the company is widening access after months of limited testing. The pitch is practical: connect your health data, then ask better questions with more context than a blank chat box ever allowed.
What can you actually do with ChatGPT Health now?
OpenAI’s health product lets users connect Apple Health and eligible medical records to ChatGPT. In return, people can view recent trends, browse their records, and try suggested prompts. That is a clear step beyond typing symptoms into a generic chatbot with no personal history attached.
Demand already looks huge. OpenAI said more than 300 million people ask ChatGPT health-related questions every week. Recent KFF Health polling cited by Mashable found that roughly one in three U.S. adults turned to a chatbot for health information in the last year.
ChatGPT Health now available is less a new habit than an attempt to organize one that already exploded. Ashley Alexander, OpenAI’s vice president of health product, framed wider access as urgent because many people cannot get quality medical care in a timely way. “Everyone should have a right to have access to this kind of support in their health,” she said.
In a media briefing, OpenAI also argued its latest models have significantly improved at health prompts, outperforming doctors in some cases. A Science study published in the spring found that OpenAI’s reasoning model o1 could diagnose real-life medical scenarios as well as or better than physicians.
How is this different from the old “ask the internet” health era?
Then: you searched symptoms, compared anonymous forum posts, and hoped a doctor visit would sort the noise. Now: ChatGPT Health is built to sit closer to your own records and wearable-style data streams. Connecting Apple Health and eligible medical records turns the chat from a general encyclopedia into a more personalized briefing tool—still imperfect, still not a clinician.
Dr. Robert Wachter, professor and chair of the Department of Medicine at the University of California, San Francisco, told Mashable he expects ChatGPT Health to help assess concerns such as the meaning of symptoms, a photo of a rash, or new lab tests or X-ray results. For severe signs—like new chest pain or weakness on one side—he said you should close the computer and see a doctor.
Wachter also noted that even highly accurate AI tools will get things wrong sometimes. The fair comparison, he argued, is not mythical perfection but the messy status quo. In his view, current tools are “far better than the status quo in many situations and are getting better quickly.”
Why does a new lawsuit matter on launch day?
Timing is awkward. A lawsuit filed Wednesday against OpenAI and CEO Sam Altman asked for an injunction that would pause healthcare-related consumer products, including ChatGPT Health. Scott Winters, a 55-year-old Florida pastor, alleged that an older ChatGPT model gave him “personalized and inaccurate” medical advice that nearly led to his death from a pulmonary embolism in 2025.
According to the filing summarized by Mashable, Winters consulted model 4o—since retired and already tied to multiple wrongful death lawsuits. Altman publicly acknowledged in April 2025 that updates to 4o made it “sycophant-y.” The complaint says ChatGPT-4o offered specific care directives without a disclaimer to consult a medical provider, and recommended treatments and medications.
When Winters reported groin tenderness—a symptom tied to the onset of his embolism—the chatbot largely dismissed the sensation. Hours later, clinicians diagnosed a pulmonary embolism from multiple blood clots in both lungs. One physician said the embolism likely occurred because Winters had been immobile; ChatGPT had recommended he limit movement to manage dizziness and blood-pressure instability.
Asked about the case, Alexander did not directly address the injunction request. She said OpenAI is “very clear” that ChatGPT is not for diagnosis and treatment. Wachter told Mashable the lawsuit involved an outdated version and that current GPT models are more trustworthy than older ones—while still warning that mistakes will happen.
How should you use ChatGPT Health without getting burned?
That caution is nostalgia in reverse: we once trusted glossy sites and anonymous boards; now the risk is a fluent assistant that sounds sure even when it is guessing.
Dr. Ashwin Ramaswamy of Mt. Sinai and Dr. Girish N. Nadkarni, chief AI officer of the Mount Sinai Health System, have studied consumer health AI responses. In a peer-reviewed advance paper in Nature Medicine earlier this year, they found ChatGPT Health’s recommendation shifted when a prompt included friends or family downplaying symptoms.
In those cases, the chatbot was 11 times more likely not to send the patient to the emergency room—even when symptoms indicated a life-threatening condition. Both physicians told Mashable the tools have real value if used carefully.
Describe what is happening concretely, and skip soft frames like “I’m probably overreacting.” Start a new conversation for each new health concern; their research under peer review linked long-running chats with answers drifting from nursing and physician triage standards.
“When in doubt, a clean conversation is the safer default,” they said. They also recommend pushing back—ask for evidence, what would make it an emergency, and what a doctor would still want to know. If ChatGPT reverses itself under mild pressure, that may signal agreeableness more than confidence.
Ramaswamy and Nadkarni also argue the field still lacks enough peer-reviewed support for company performance claims and called on OpenAI to open products to independent evaluation.
Bottom line: ChatGPT Health now available expands a tool people already lean on, with powerful data hooks and a reminder that fluent answers are not clinical care. Use the connections carefully, and treat urgent symptoms as a reason to see a human doctor.