ChatGPT connects to your medical records and Apple Health

Most people don't immediately call a doctor when they get confusing blood work back. They either paste the numbers into a search engine or dump the entire lab report into a chatbot.

The annoying part has always been context. Every time you opened a new chat, you had to re-explain your medications, past lab results, and how poorly you slept this week. OpenAI wants to eliminate that friction.

OpenAI rolled out Health in ChatGPT to logged-in U.S. users aged 18 and older. It spans web and iOS across Free, Go, Plus, and Pro tiers, letting users securely sync medical history directly into their everyday AI assistant.

Health questions happen in everyday chats

OpenAI notes that more than 300 million people now bring health-related questions to ChatGPT every single week, up from 230 million when early testing began in January.

Back then, the company built a dedicated health hub where users had to navigate to a separate tab to talk about medical issues. Early testing revealed an obvious flaw: over 70% of health discussions happened completely outside that silo.

People planning dinner recipes naturally wanted ChatGPT to consider their food allergies. Families mapping out a weekend trip asked for low-impact activities because of a recent ankle injury. Forcing them into an isolated medical space was just an unnecessary barrier.

Under the new setup, Health lives in the left sidebar as your control center, but the context flows across normal chats. Users can connect Apple Health alongside supported medical records from hospital systems like Epic and Oracle Health, plus digital platforms like One Medical and Function Health. Wearables and fitness apps like MyFitnessPal or Weight Watchers can also funnel activity and sleep data through Apple Health.

When you ask a question anywhere in ChatGPT—or tag @Health explicitly—the model can pull relevant context with your permission. Free accounts run on GPT-5.5 Instant, which OpenAI claims matches earlier reasoning models on tough health evaluations, while paid tiers get GPT-5.6 Sol, engineered with input from hundreds of global physicians.

Big claims meet an awkward lawsuit

Handing over intimate medical data to an AI company naturally sets off alarm bells, so OpenAI put several guardrails in place.

Synced health records, Apple Health data, and the conversations utilizing them will never be used to train frontier models or target ads, regardless of your standard data-training settings. Data is encrypted in transit and at rest, and disconnecting an account wipes the synced data from OpenAI's servers within 30 days. By default, ChatGPT prompts you for permission each time it references your records, though you can toggle on "always allow."

Yet the launch was overshadowed by internal hype and external legal trouble.

During a press briefing, Ashley Alexander, OpenAI's VP of health product, claimed the company's models "are now capable of reasoning at levels that are better than clinician level." That forced OpenAI health lead Karan Singhal to step in and publicly temper the statement, cautioning against sweeping generalizations even as select academic papers show promise.

To make matters more tense, a Florida pastor filed a lawsuit against OpenAI just one day before the rollout. The suit alleges ChatGPT gave him near-fatal guidance suggesting he put off urgent medical care for what turned out to be a pulmonary embolism. In response, OpenAI leaned hard on its legal disclaimers, reiterating that ChatGPT Health is explicitly not designed to diagnose or treat conditions.

Why it matters

Breaking health context out of an isolated corner and injecting it into everyday conversation is a very smart product move.

For users, it fixes the single biggest headache in personal health tracking: fragmentation. Instead of your cholesterol labs living in one patient portal, your sleep metrics trapped in an Apple Watch, and your daily meal planning sitting in ChatGPT, the AI can finally connect the dots. If you are prepping for an annual checkup or trying to decipher why a lab value shifted over three years, this is genuinely useful.

At the same time, this convenience introduces serious behavioral risk.

When a chatbot has access to your full clinical background and speaks with polished authority, human nature makes it nearly impossible not to treat it like a primary care doctor. Having an executive boast that AI reasons better than clinicians—right as a lawsuit lands alleging delayed emergency care—highlights the tightrope OpenAI is walking.

Competitors like Google and Anthropic are circling the same healthcare territory with immense caution. We think ChatGPT Health is worth trying if you need help decoding medical jargon or framing smart questions for your next appointment. But keeping the final diagnostic call firmly in the hands of a licensed physician remains non-negotiable.