On Thursday, OpenAI announced that its health‑focused chatbot, ChatGPT Health, will be available to any adult in the United States. The service will tap into users’ medical records and Apple Health data, allowing people 18 and older to ask for guidance on diet, exercise, and other wellness topics.
The rollout follows a period of heightened scrutiny for OpenAI. In the last three months, two lawsuits have accused the company of providing dangerous medical advice through its AI. OpenAI says the new feature is designed with stricter safety layers and aims to give personalized, evidence‑based recommendations while complying with privacy regulations.
ChatGPT Health will request permission to read electronic health records stored in compatible systems and data from Apple’s Health app. Once granted, the model can generate suggestions tailored to an individual’s medical history, allergies, and current medications. OpenAI claims the AI has been trained on up‑to‑date clinical guidelines and will flag any response that falls outside safe parameters. Users can opt out of data sharing at any time, and the company promises to delete stored information after the session ends. Early testers reported smoother interactions compared with earlier beta versions, noting clearer explanations and more cautious language.
Legal experts warn that the lawsuits could shape how the service operates. One case alleges that previous versions of the chatbot suggested risky treatments, leading to severe health outcomes. A second suit claims the AI failed to warn users about potential side effects of over‑the‑counter medications. OpenAI’s legal team argues that the new version incorporates real‑time risk assessments and that users are required to acknowledge a disclaimer before each query. Regulators are watching closely, and any future court rulings may force the company to limit the scope of medical advice or enhance oversight mechanisms.
The nationwide launch signals OpenAI’s confidence in its safety upgrades, but the company remains vulnerable to further legal action. If courts find the AI’s advice negligent, OpenAI could face hefty penalties and be forced to retract the feature. Conversely, successful defense could set a precedent for broader AI‑driven health tools. The industry will monitor user adoption rates and any emerging safety incidents to gauge the model’s long‑term viability.
What data does ChatGPT Health access? The service can read electronic health records from participating providers and pull metrics from Apple Health, such as heart rate and activity logs, only after user consent.
Is the advice medically certified? Responses are generated from large language models trained on medical literature, but they are not a substitute for professional diagnosis. Users are urged to confirm suggestions with a qualified healthcare provider.
How can users protect their privacy? OpenAI allows users to revoke data permissions at any time and deletes session data after each interaction, aiming to meet U. S. privacy standards.