OpenAI has introduced ChatGPT Health, a new dedicated experience within ChatGPT designed to help people better understand and manage their health and wellness using their own medical and lifestyle data—while maintaining strong privacy and security protections.

Health is already one of the most common ways people use ChatGPT. According to OpenAI’s de-identified analysis, more than 230 million people globally ask health and wellness questions on ChatGPT each week. ChatGPT Health builds on this usage by offering a separate, purpose-built environment that securely connects health information with AI-powered assistance—supporting, but not replacing, care from medical professionals.

Health information today is often fragmented across portals, apps, wearables, PDFs, and medical notes, making it difficult for individuals to see the full picture of their health. Many people already turn to ChatGPT to help make sense of test results, symptoms, and next steps.

ChatGPT Health brings these conversations into a dedicated space where responses can be informed by connected health data and broader personal context. Users can securely link medical records and wellness apps—such as Apple Health, Function, and MyFitnessPal—to receive more relevant, personalized guidance.

The experience is intended to help users:

Health is explicitly not intended for diagnosis or treatment. Instead, it focuses on everyday understanding, long-term patterns, and better preparation for medical conversations with clinicians.

Given the sensitivity of health data, ChatGPT Health operates as a separate space within ChatGPT, with enhanced privacy protections and security controls.

Health lives in its own dedicated environment, with:

While Health chats appear in your ChatGPT chat history for easy access, the health information itself remains contained within the dedicated Health space. ChatGPT may occasionally use limited non-Health context—such as a recent move or lifestyle change—to make a health conversation more relevant; however, Health data never flows back into standard ChatGPT chats, and conversations outside of Health cannot access files, conversations, or memories created within Health.

OpenAI notes that several privacy and data controls already exist across ChatGPT, providing users with meaningful control over their data.

Foundational protections across ChatGPT include:

For ChatGPT Health, these protections are augmented with additional, health-specific safeguards, including:

Users can also enable multi-factor authentication (MFA) to add an extra layer of account protection.

When users choose to connect their health data, responses in Health are grounded in their own information.

To enable access to trusted U.S. healthcare providers, OpenAI partners with b.well, the largest and most secure network of live, connected health data for U.S. consumers, which adheres to the highest industry standards in data security and privacy.

Users can upload files directly or connect medical records and apps through tools (+) or the “Apps” section of Settings.

Supported integrations include:

All apps included in Health must meet OpenAI’s privacy and security requirements, collect only the minimum data necessary, and undergo additional security review specific to inclusion in Health. When an app is connected for the first time, users are shown what types of data may be collected by the third party, and access can be disconnected at any time, immediately revoking access.

App connections require explicit user permission—even if an app is already connected to ChatGPT for conversations outside of Health—and users can revoke access to connected apps or medical records at any time through the “Apps” section of Settings.

ChatGPT Health was developed in close collaboration with physicians worldwide to ensure responses are helpful, safe, and clinically appropriate.

Over two years, more than 260 physicians across 60 countries and dozens of specialties provided feedback on model outputs—contributing over 600,000 evaluations across 30 focus areas. This physician feedback shaped not only what Health can do, but how it responds—guiding when to encourage follow-ups with clinicians, the urgency of those recommendations, how to communicate clearly without oversimplifying, and how to prioritize safety in moments that matter.

This physician-led approach is embedded in the model powering Health and evaluated using HealthBench, an assessment framework OpenAI created with physician input. HealthBench evaluates responses using physician-written rubrics that reflect how clinicians judge quality in practice, prioritizing safety, clarity, appropriate escalation of care, and respect for individual context.

This evaluation-driven approach supports practical tasks people already seek help with, such as explaining lab results, preparing appointment questions, interpreting wearable data, and summarizing care instructions. The result is assistance people can trust, designed to complement the role of healthcare providers rather than replace them.

Users can sign up for the ChatGPT Health waitlist as access rolls out. Initial availability is limited to a small group of early users to refine the experience.

Once access is granted, users can select Health from the ChatGPT sidebar, connect their medical records and apps, upload files, and begin health-focused conversations using text, voice, search, and dictation.

Users can also add custom instructions within Health to guide what ChatGPT focuses on, avoid certain sensitive topics, or adjust how responses are framed. These instructions apply only to Health conversations and can be updated or removed at any time through Health or Settings.

Q: Is ChatGPT Health intended to diagnose or treat medical conditions?
A: No. ChatGPT Health is not designed for diagnosis or treatment. It is intended to help users understand health information, recognize patterns over time, and prepare for conversations with healthcare providers—while explicitly supporting, not replacing, clinical care.

Q: Are conversations in ChatGPT Health used to train OpenAI’s foundation models?
A: No. Conversations within ChatGPT Health are not used to train OpenAI’s foundation models, reflecting the added privacy protections built specifically for sensitive health information.

Q: How is health data kept separate from regular ChatGPT conversations?
A: ChatGPT Health operates as a dedicated space within ChatGPT, with separate conversations, memories, connected apps, and files. Health data does not flow back into non-Health chats, and standard ChatGPT conversations cannot access Health information.

Q: Can users control or remove connected health data and memories?
A: Yes. Users can view, manage, or delete Health memories at any time and disconnect medical records or wellness apps instantly through Settings, which immediately removes access.

ChatGPT Health reflects a growing effort to make AI more useful in everyday life—particularly in complex, high-stakes areas like healthcare—while maintaining clear boundaries.

It illustrates how AI tools may begin to handle sensitive domains like health differently from general-purpose chatbots. By separating health conversations into a dedicated, privacy-protected space and grounding responses in user-approved data, OpenAI is signaling that not all AI interactions should be treated the same. Rather than positioning AI as a diagnostic tool, ChatGPT Health is framed as support for understanding, preparation, and ongoing context—without replacing the role of clinicians.

As access expands, ChatGPT Health may offer a clearer example of how trust in AI is built: through explicit limits, physician-guided design, and security choices that prioritize safety alongside personalization. In practice, this approach could support long-term wellness awareness, more informed decision-making, and more productive conversations between patients and healthcare providers.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.