In an age when our most intimate data—our bloodwork, our steps, our sleep—lives scattered across a dozen disconnected platforms, OpenAI has taken a deliberate step toward consolidation, launching ChatGPT Health as a dedicated space where personal medical records and wellness data can be gathered into a single, conversational relationship. The feature, rolling out gradually to select users in early 2026, is framed not as a replacement for physicians but as a kind of informed companion—one that helps people arrive at the doctor's office already understanding their own story. It is a quiet but co
OpenAI launches ChatGPT Health with secure medical records integration
Health data stays isolated, encrypted, excluded from model training
So OpenAI is letting people upload their actual medical records into ChatGPT now? That seems like a significant shift.
Yes, but it's important to understand what they're doing with it. The records stay in an isolated environment—they don't feed into the training of OpenAI's main models. It's a separate space, encrypted, with its own security layer.
How do we know that isolation actually works? OpenAI says the data is excluded from training, but we're trusting their architecture here. Has this been independently audited?
That's a fair question. The company says it's been developed with physician input and includes privacy safeguards, but you're right that we don't have third-party verification yet.
What can someone actually do with this? What's the practical use case?
Let's say you get lab results and they're confusing. You can upload them, and ChatGPT can help you understand what the numbers mean in context of your health history. Or you're preparing for a doctor's appointment and want to organize your questions. It pulls from your fitness data, your diet tracking, your medical history.
But it's not giving medical advice, right? OpenAI was very clear about that.
Correct. It's framed as a tool to help you understand your information better, not as a replacement for a doctor. The company is being careful about that distinction.
Who can actually use this right now?
It's in early access—a small group of users on free and paid plans, but not in Europe, the UK, or Switzerland. Some features like Apple Health integration only work on iOS. Wider rollout is coming in the coming weeks.
So we're looking at a gradual, limited launch. That's smart from a liability perspective, but it also means most people can't test whether this actually works as promised.
Le Pouls
- Health questions are already among the most common things people ask ChatGPT, yet until now the AI had no access to the personal data that would make those answers truly meaningful.
- Medical information remains maddeningly fragmented—trapped in hospital portals, PDF attachments, wearable apps, and nutrition trackers that never speak to one another—creating a gap that ChatGPT Health is explicitly designed to bridge.
- Privacy concerns loom large: OpenAI has responded by isolating health conversations in an encrypted environment entirely separate from a user's other chats, and has committed to excluding all health data from AI model training.
- The rollout is deliberately cautious—limited to a small group of users, unavailable in the EU, Switzerland, and the UK, and dependent on iOS for certain integrations—signaling that OpenAI is moving carefully in a domain where trust is everything.
- Amazon is simultaneously testing its own Health AI assistant, but with a commerce-oriented lens, revealing that the race to define AI's role in healthcare is already producing competing visions of what that relationship should look like.
In an age when our most intimate data—our bloodwork, our steps, our sleep—lives scattered across a dozen disconnected platforms, OpenAI has taken a deliberate step toward consolidation, launching ChatGPT Health as a dedicated space where personal medical records and wellness data can be gathered into a single, conversational relationship. The feature, rolling out gradually to select users in early 2026, is framed not as a replacement for physicians but as a kind of informed companion—one that helps people arrive at the doctor's office already understanding their own story. It is a quiet but consequential moment in the long negotiation between human health, personal privacy, and the expanding reach of artificial intelligence.
OpenAI has launched ChatGPT Health, a dedicated section of its platform built specifically for health and wellness conversations. Users can connect wellness apps—Apple Health, MyFitnessPal, Peloton, and others—and upload medical records, lab results, and files, allowing the AI to draw on their actual personal data rather than offering only generic guidance. The goal is to help people make sense of their own health information: understanding bloodwork, tracking fitness trends over time, and arriving at medical appointments better prepared.
The company developed the feature with physician input and has been clear that it is not intended to replace professional care. Instead, it positions ChatGPT Health as a tool for informed self-understanding—a way to consolidate the health data that typically sits scattered across hospital portals, apps, and devices into a single, coherent conversation.
Privacy is built into the architecture. The Health space operates in an isolated environment, with conversations and uploaded files kept entirely separate from a user's other ChatGPT activity. Health data is excluded from training OpenAI's models, encrypted at rest and in transit, and can be viewed or deleted by the user at any time. Multi-factor authentication is available for additional security.
The initial rollout is limited—available to a small group across free and paid plans, but not yet in the European Economic Area, Switzerland, or the UK. Broader availability on web and iOS is planned in the coming weeks, with more integrations to follow.
The launch arrives as Amazon tests its own Health AI assistant, one oriented more toward general wellness information and its pharmacy ecosystem. The two approaches reveal a divergence in strategy: OpenAI is betting on deep, private integration with personal medical data, while Amazon leans toward informational guidance tied to its existing health services. Together, they mark an early chapter in what will likely be a long and consequential negotiation over AI's place in human healthcare.
OpenAI has introduced ChatGPT Health, a new section of its ChatGPT platform designed specifically for health and wellness conversations. The feature allows users to upload medical records and connect to wellness applications—Apple Health, MyFitnessPal, Weight Watchers, Peloton, and others—so that when they ask health-related questions, the AI can draw on their personal data rather than offering only general information.
The company says health inquiries are already among ChatGPT's most frequent use cases. ChatGPT Health formalizes this by creating a dedicated space where a user's scattered health information—typically spread across hospital portals, apps, PDFs, and wearable devices—can be consolidated and referenced in conversation. Users can upload files, take photos of lab results, use voice input, and ask ChatGPT to help them understand bloodwork, track fitness trends, prepare questions for doctor appointments, or review their diet and exercise patterns over time. The system can identify patterns across time and summarize information in ways that might help someone feel more informed before speaking with a healthcare provider.
OpenAI developed the feature with input from physicians and has been explicit that ChatGPT Health is not meant to replace medical professionals. Instead, the company frames it as a tool to help people better understand their own health information and feel more prepared during medical visits. Users can set custom instructions specific to the Health space to control how information is presented or to avoid certain topics.
The rollout is limited for now. Access is available only to a small group of users across ChatGPT's free, Go, Plus, and Pro plans, and only outside the European Economic Area, Switzerland, and the UK. Some integrations, including Apple Health support, require an iOS device. OpenAI plans to expand availability on web and iOS in the coming weeks and to add more supported apps and capabilities based on user feedback.
Privacy is central to the design. ChatGPT Health operates in an isolated environment, keeping health conversations, uploaded files, and memories separate from a user's other ChatGPT chats. Health data cannot be accessed outside the Health space and is explicitly excluded from training OpenAI's foundation models. All data is encrypted both at rest and in transit, with additional encryption layers applied within the Health environment. Users can view or delete their Health memories at any time and can enable multi-factor authentication for extra security.
The launch comes as other tech companies move into health AI. Amazon has been testing its own health assistant, called Health AI, which answers general wellness questions and recommends products. Amazon's approach appears more informational and tied to its pharmacy and medical services offerings, whereas ChatGPT Health emphasizes secure integration of personal medical data within a privacy-protected space. The two services represent different strategies for bringing AI into healthcare—one focused on helping individuals understand their own records, the other on guiding users toward health information and services.
Citations marquantes
ChatGPT Health is designed to help users better understand their information and feel more prepared when speaking with healthcare providers, not to replace medical professionals— OpenAI