On July 23, 2026, OpenAI launched ChatGPT Health for all adult users in the US, integrating Apple Health data and medical records, but the launch is accompanied by a legal crisis over the chatbot's deadly and dangerous health advice.
On Thursday, July 23, 2026, OpenAI officially launched ChatGPT Health for all registered users over 18 in the United States, on both web and iOS platforms. The feature, which had been in beta since January 2026, is now available to users on all plans, from free to Pro tier, allowing the chatbot access to medical records and Apple Health data with prior user consent.
According to an OpenAI press release, ChatGPT Health can compare lab results, summarize changes since the last doctor's visit, track medication intake, and analyze how sleep, physical activity, and workouts affect the user's health. A key new feature compared to the beta version is that health data is not separated into a dedicated space within the app but is directly integrated into the conversational interface. This means that if the user consents, the chatbot will consider health data even for questions that are not seemingly medical. As OpenAI stated in the release:
'ChatGPT can take into account a dietary restriction when helping you choose a restaurant, or a recent injury when planning low-intensity weekend activities with the family.'
OpenAI emphasizes that the connected medical records and Apple Health data are not used for model training or targeted advertising, nor are conversations that include health data. Users who do not want to link medical records can still use the health section for general health questions. Nevertheless, OpenAI acknowledges in the release that the chatbot 'can still make mistakes' and urges users to 'verify important information and discuss medical decisions with your doctor.'
The launch comes as OpenAI faces two high-profile lawsuits directly related to the chatbot's health advice. In a civil lawsuit filed in May 2026, the family of 19-year-old Sam Nelson claims that the chatbot advised the young man to use Xanax to 'mitigate' the effects of Kratom, an herbal supplement with opioid-like effects. This combination allegedly led to an overdose and Nelson's death in May 2025. The family, asserting that Nelson 'died following medical advice that ChatGPT was programmed to provide,' seeks not only financial compensation but also a temporary suspension of the ChatGPT Health feature. The incident reportedly involved the GPT-4o model.
Just one day before the official launch, on July 22, 2026, a pastor from Florida also filed a lawsuit against OpenAI. According to the complaint, instead of directing him to seek emergency medical help for early symptoms of a life-threatening pulmonary embolism, the chatbot responded that 'God did not design your body to fail endlessly,' thereby allegedly delaying urgent treatment. This lawsuit also seeks a temporary suspension of the ChatGPT Health feature, financial compensation, and stricter regulatory oversight of the chatbot's ability to answer health-related questions.
According to OpenAI, more than 300 million users weekly use ChatGPT for health-related questions, illustrating how widespread reliance on AI in medicine already is, and all without any formal regulatory oversight. For European users, including those in Croatia, ChatGPT Health is currently not available, but that could change as the service expands. The European Union is already applying parts of the EU AI Act, adopted in 2024, which imposes strict requirements on AI systems in the medical field classified as high-risk. The question of when and under what conditions ChatGPT Health might reach the European market remains open, and the ongoing US lawsuits could also influence the regulatory framework on this side of the Atlantic.
