OpenAI has begun rolling out ChatGPT Health across the United States, giving eligible adults the ability to plug their medical records and fitness data directly into the world's most widely used AI chatbot. The launch, which started July 23, marks one of the most ambitious attempts yet to put artificial intelligence at the center of everyday personal health management.
The scale of the opportunity helps explain the move.
According to OpenAI, more than 300 million people already turn to ChatGPT with health-related questions every week. They ask it to decode lab results and prepare for doctor visits. They ask it to make sense of medical jargon. Until now, those conversations relied on generic information. With Health, the chatbot can ground its answers in a user's own data: blood tests, visit summaries, medication lists, sleep patterns, and workout history.
What's Actually New
The feature is available to logged-in users aged 18 and older in the US, on both web and iOS, and spans every consumer tier: Free, Go, Plus, and Pro. It appears as a dedicated Health section in the ChatGPT sidebar, where users are guided through connecting two broad categories of data.
The first is Apple Health, linked through Apple's HealthKit framework, which brings in activity and sleep metrics along with fitness data. The second is formal medical records from major hospital and provider networks, including systems built on the two largest electronic health record platforms in the country, as well as popular primary care and diagnostics services.
Once connected, the assistant can compare a new lab result against prior tests, summarize what has changed since a last appointment, track prescriptions over time, and explore how sleep and exercise habits relate to broader health trends. The context also flows into ordinary conversations: the chatbot might factor in a dietary restriction when suggesting restaurants, or account for a recent injury when helping plan a weekend hike.
A Second Attempt, Backed by Big Numbers
This is not OpenAI's first swing at the idea. The company piloted a health experience in January 2026 as a separate, walled-off portal — and the results reshaped the product. Testing reportedly showed that roughly 70 percent of health questions surfaced naturally inside routine chats rather than in a dedicated health tab, prompting the company to weave health context throughout the main experience while keeping sensitive records stored in their own protected space.
The medical rigor behind the relaunch is where the numbers get striking. OpenAI says it worked with a network of more than 260 practicing physicians across roughly 60 countries and dozens of specialties. Those doctors reviewed and scored model outputs over 600,000 times across 30 health domains. That feedback trained the system to prioritize safety and clear communication, and to recognize when a user should be pushed toward professional care. The physician network also helped build a new evaluation framework that grades health answers against doctor-written rubrics rather than multiple-choice exam questions.
Under the hood, responses are powered by the company's newest models, with paid subscribers getting its most capable reasoning system and free users a faster variant.
The Privacy Question
OpenAI is leaning hard on privacy assurances. Connected medical records and Apple Health data are stored separately from regular chats, are not used to train the company's foundation models, and are not used for ad targeting. The chatbot asks permission before drawing on health data to personalize any response, and the Health section works even for users who choose not to connect anything.
Still, a caveat looms.
Data shared with a consumer chatbot is generally not covered by HIPAA, the federal law that protects health information held by doctors and insurers. Protection here rests on company policy rather than statute, a distinction privacy advocates are quick to highlight.
OpenAI is equally careful about the medical fine print. Its terms state the service is not intended for diagnosing or treating any condition, and the company frames Health as a tool for understanding information and becoming a better-informed patient, not a replacement for clinical care. Independent research has repeatedly cautioned that AI chatbots remain unreliable as sources of direct medical advice.
The Bigger Picture
The consumer launch caps a rapid push into healthcare. In April, the company released a free documentation and research tool for verified physicians, and it has stood up an enterprise-facing healthcare initiative aimed at hospital systems. Rivals in the AI race have shipped their own health features, making personal health one of the industry's most hotly contested frontiers.
For now, the rollout is limited to the US, with no timeline announced for international expansion. The feature is also absent from the company's coding environment.
Comments