AI Products
OpenAI Launches Health in ChatGPT
OpenAI has begun a U.S. rollout of Health in ChatGPT, enabling people to connect eligible health information and use it in health-focused conversations with privacy controls.
What OpenAI announced
OpenAI has started a U.S. rollout of Health in ChatGPT. The new experience is intended to help people bring eligible health information into conversations so they can better understand and navigate their own care. OpenAI describes the feature as a way to connect health information, review relevant context, and ask questions in a health-focused space inside ChatGPT.
The announcement matters because health is a more sensitive use case than ordinary productivity chat. A useful response may depend on personal context, but that same context deserves tighter controls than a casual prompt. Availability, eligible connections, and product behavior can vary by account, region, and connected service, so users should check the settings shown in their own product before relying on a workflow.
Privacy and boundaries come first
Before connecting a provider, wearable, or health app, users should understand exactly what is being connected, which data may be included in a conversation, and how to review or remove that context later. It is sensible to keep health conversations focused, avoid sharing information that is not necessary for the question, and revisit permissions when a device, provider, or care situation changes.
The feature can help organize questions, explain unfamiliar terminology, and surface topics to discuss with a professional. It should not be treated as a diagnostic service, a replacement for emergency care, or an authority that can settle a clinical decision. Symptoms that are urgent, worsening, or potentially dangerous need appropriate medical attention rather than a chat workflow.
A practical way to use it
Start with a low-risk task: prepare questions for an appointment, summarize information you already have, or identify plain-language terms you want a clinician to explain. Keep the original records available, note where a response may be uncertain, and verify medication, diagnosis, and treatment decisions with a qualified professional.
For teams building health-adjacent workflows, the launch is a reminder to define the purpose narrowly. Good safeguards include clear consent, minimum necessary data, human escalation for high-impact questions, and a record of what information was used. The important test is not whether a chat answer sounds fluent; it is whether the workflow preserves privacy, communicates uncertainty, and helps a person make a safer next step.