OpenAI has begun rolling out Health in ChatGPT to logged-in users 18 and older in the U.S., allowing people to connect Apple Health and supported medical-record sources so ChatGPT can use that information to ground health conversations. The feature aims to help users compare results, track changes over time and bring wearable or medical-record context into everyday chats while applying explicit permission prompts and layered privacy protections.
OpenAI launches Health in ChatGPT for U.S. users
OpenAI said Health is available starting today on web and iOS across Free, Go, Plus and Pro plans for users aged 18 and up in the United States. The product lets users choose to connect Apple Health and supported medical records—including certain U.S. hospital systems, One Medical, and Function Health—so ChatGPT can reference relevant medications, lab results, visits, sleep and activity when answering questions.
OpenAI frames the feature as a response to user demand: the company says more than 300 million people turn to ChatGPT weekly with health-related questions. It also shared findings from early testing that more than 70% of health-related conversations happened outside a separate health workspace, prompting the company to let connected health information be available across regular chats with user permission.
How Health accesses and applies connected data
Users opt in to connect accounts from the Health pane in ChatGPT’s sidebar or More menu. Once a source is linked, data begins syncing and may take a few minutes to appear. While syncing, users can add other accounts or return later. If a wearable or third-party app shares data to Apple Health, that data can flow into ChatGPT when Apple Health is connected and the user permits access; availability depends on what each app exposes through Apple Health.
By default, ChatGPT will prompt for permission before using connected medical records or Apple Health information in a response. Users can approve each request or set a preference to always allow access. Users can also explicitly request health context by adding @Health to a message. Accounts can be disconnected at any time; OpenAI says data synced from a disconnected source is deleted from OpenAI’s systems within 30 days, although any health information already included in conversation history remains until the user deletes those conversations.
OpenAI also describes controls around memory: Memories can be created from Health conversations to personalize future chats, but memories are not created directly from connected medical records or Apple Health information. For actions that might disclose health information through other connected plugins—such as sending a training plan based on Apple Health data—OpenAI says additional safeguards will check that the sharing matches the user’s request and may require confirmation.
Why it matters
The update aims to reduce the friction of repeatedly telling ChatGPT the same health details or manually aggregating scattered records. With connected context, ChatGPT can compare a new lab result with prior tests, summarize changes since a last appointment, or relate sleep and activity trends to a user’s routine, potentially helping people prepare questions for clinicians or manage day-to-day choices with more tailored information.
OpenAI also emphasizes model-level improvements for health: GPT‑5.5 Instant is available to free users and was previously credited with gains in recognizing urgent-care signals and asking for clarifying context. GPT‑5.6 Sol, available to paid users, is presented as stronger on tasks requiring reasoning across complex health details. OpenAI reports physician-led evaluations and a HealthBench Professional metric where GPT‑5.6 outperformed GPT‑5.5 on challenging health tasks.
Privacy, safety and model-training limits
OpenAI positions Health with layered safeguards. It states explicitly that information connected in Health and conversations that use it are not used to train OpenAI’s foundation models or to target ads. Conversations that do not use connected Health data continue to follow the user’s standard ChatGPT model-training settings, which can be changed under Settings > Data Controls.
Physicians were involved in developing evaluation rubrics and tested Health in ChatGPT before release, but OpenAI also cautions that the system can still make mistakes and is not a replacement for qualified medical judgment. The company encourages verification of important information with healthcare providers.
What to watch next
OpenAI says Health is beginning its rollout; the company does not specify how quickly availability will expand beyond the initial U.S. web and iOS launch or whether additional medical-record partners will be added. Other open questions include how connected-data policies will evolve, whether enterprise or API integrations for health contexts will follow, and how longitudinal evaluations by physicians will be published or shared.
Users should also watch the controls that govern permission prompts, memory creation and deletion: OpenAI’s 30-day deletion promise for synced account data leaves conversation history intact until manually removed, which may affect how long connected details persist in the product.
Health in ChatGPT brings personal health records and wearable data directly into everyday chats with model improvements and explicit privacy settings, but OpenAI says users should verify important medical information with a qualified professional and manage connected data and conversation history through the product’s controls.
Source: OpenAI