OpenAI has announced that it has officially made its controversial health assistant feature ChatGPT Health available to all users in the United States. Users aged 18 and over can share their medical records and Apple Health data with the chatbot and get advice on diet, exercise routines or medical test results.

This feature, which was previously tested with a limited group of users, was developed in line with the feedback and integrated directly into the chat interface. If users give their consent, their health information can be used as context not only in health-related questions but in all interactions with the chatbot.

New features and technical infrastructure

OpenAI noted that the ChatGPT Health feature now works with GPT-5.6 Sol, the company’s current flagship model, for premium users. The company claims that this new model is the most powerful option they have ever developed in terms of health. On the other hand, it should be reminding that the GPT-5.6 Sol model had previously come to the fore with claims that it had carried out a cyber attack against HuggingFace.

OpenAI states that the system can offer suggestions on topics such as choosing a restaurant or planning family activities, taking into account information such as dietary restrictions or a recent injury. Users can detail the responses given by the chatbot with follow-up questions or correct missing information.

Legal processes and security concerns

This large-scale rollout decision comes in the shadow of significant legal issues that OpenAI faces due to its health advisories. In a lawsuit filed in May, the family of a 19-year-old teenager named Sam Nelson sued the company, claiming that he died as a result of the advice given by ChatGPT.

The family demands that the ChatGPT Health feature be stopped, arguing that the chatbot suggested taking Xanax after using Kratom and that this led to death.

Similarly, a pastor living in Florida also sued the company, stating that ChatGPT delayed seeking emergency medical assistance for a pulmonary embolism diagnosis due to health recommendations. The priest claims that the chatbot suggested him in religious terms rather than telling him that he needed medical help.

OpenAI states that more than 300 million people a week use ChatGPT to ask health-related questions, but acknowledges that the system can have flaws. The company recommends that users verify valuable information and definitely discuss their medical decisions with a healthcare professional.

Do you think artificial intelligence-based health assistants are sufficiently confident in giving medical advice?