OpenAI is rolling out “Health in ChatGPT” for users in the United States aged 18 and older, allowing the conversational AI to securely connect with Apple Health and supported electronic medical records. According to OpenAI, the feature enables the platform to reference lab results, medications, and fitness data during chats, aiming to reduce repetitive data entry for the more than 300 million people who use the chatbot for health questions each week.
How Health in ChatGPT Works and Models Behind It
With explicit user permission, ChatGPT can access connected information including medications, lab results, recent clinical visits, sleep metrics, and daily activity. According to OpenAI, the tool can compare new lab results against prior tests, summarize changes since a patient’s last appointment, or connect sleep and workout data to an established routine. The feature is designed to support, not replace, care from medical professionals. Users can connect Apple Health, supported medical records from U.S. hospital systems, One Medical, or Function Health by opening the Health section from the sidebar or More menu, selecting “Get started,” and completing the guided connection process. To ensure accuracy, OpenAI states that it collaborates with hundreds of physicians to develop health scenarios and assessment rubrics covering safety, communication, context awareness, and clinical escalation.
Underpinning these conversations are two distinct models deployed by OpenAI. The Free plan utilizes GPT-5.5 Instant, which company documentation describes as better at recognizing when urgent care may be needed, requesting relevant context, explaining uncertainty, and simplifying complex information. Paid users gain access to GPT-5.6 Sol, which OpenAI categorizes as its most advanced health model with improved performance on tasks requiring complex reasoning. According to internal benchmarks released by OpenAI, all GPT-5.6 models outperformed GPT-5.5 on HealthBench Professional, an evaluation designed for challenging health tasks.
Privacy Safeguards and Data Controls
To address data privacy concerns, OpenAI states that Health is built with layered safeguards, ensuring all conversations are encrypted at rest and in transit, with connected medical records receiving additional encryption. According to OpenAI, connected health information, Apple Health data, and the conversations utilizing them are entirely excluded from foundation model training and ad targeting, regardless of a user’s general model-training settings. Conversations that do not access health data follow standard model-training preferences, which users can adjust under Settings and Data Controls. By default, ChatGPT prompts the user for permission before utilizing connected records to personalize a response. Users can choose to approve each request individually or enable always-on access by modifying settings under Plugins and Health. Furthermore, users can disconnect their health data at any time through the Account menu; OpenAI states that disconnected data is purged from its systems within 30 days, although records already printed into individual chat histories remain until those specific conversations are deleted.
Availability and Limitations
The feature is available on web and iOS across Free, Go, Plus, and Pro tiers for logged-in U.S. users, accessible via the ChatGPT sidebar. OpenAI has not yet announced an Android release timeline, support in Codex, or international availability outside the United States. Because connected records may occasionally be incomplete or outdated, OpenAI emphasizes that ChatGPT does not replace qualified medical professionals and encourages users to verify important clinical details with their healthcare providers or the original data source.
Worth a look