OpenAI announced Health in ChatGPT, a new feature enabling eligible U.S. users to securely connect their electronic health records and Apple Health data directly to the chatbot for personalized health insights. The feature allows patients to upload medical documents and ask ChatGPT questions about their health data, potentially surfacing patterns or prompting conversations about their conditions. While positioned as a way to help users "better understand their health," the launch highlights a critical regulatory gap: ChatGPT itself is not FDA-cleared as a medical device, and OpenAI has not published explicit liability frameworks for instances where the AI misinterprets, misrepresents, or makes errors in analyzing sensitive patient information. The company's terms of service already disclaim medical advice, but the nature of this integration—where users are directly querying an LLM about their actual medical records—blurs the line between a conversational tool and a medical assistant in ways regulators have not yet addressed.
The feature integrates with major EHR systems and Apple Health, creating a direct pipeline from patient records to OpenAI's servers. Critically, OpenAI has not disclosed which specific EHR vendors are supported at launch, the exact data retention policies for uploaded medical documents, or whether users can request deletion of health data from OpenAI's systems after analysis. Healthcare privacy experts have flagged this as a regulatory tension: while HIPAA technically allows patients to share their own data with third parties, it does not require those third parties to implement HIPAA-level safeguards unless they are designated "business associates." OpenAI's legal status vis-à-vis HIPAA compliance remains unclear. Additionally, there is no indication that Health in ChatGPT has undergone FDA premarket review, unlike competitor tools from Google (Google Health) and Amazon (HealthLake), which have pursued clearer regulatory pathways. The business model—whether this remains free-tier or moves to premium subscriptions—has not been specified, raising questions about whether OpenAI intends this as a growth driver for paid ChatGPT accounts.
The launch arrives amid growing competitive pressure in AI-assisted healthcare and reflects OpenAI's expansion into high-stakes verticals. However, the absence of disclosed safety validation, liability indemnification language, or regulatory pre-clearance suggests OpenAI is moving faster than existing medical AI precedent. Healthcare lawyers and privacy advocates have called for clarity on who assumes liability if ChatGPT makes a harmful error based on a patient's medical records, and whether OpenAI plans to seek FDA clearance or pursue a regulatory safe harbor. Until those questions are answered, Health in ChatGPT exists in legal and regulatory limbo—available to millions of patients, but with governance frameworks that lag the stakes of the data being processed.