OpenAI has updated the rules for using its services, and from now on, ChatGPT is prohibited from providing personalized medical or legal advice.
According to the new Terms of Use, starting October 29, 2025, ChatGPT cannot act as a “consultant” on matters requiring a license—specifically in medicine, law, or finance. Instead, the bot can still provide general information, explain terms or processes, and show examples, but it will direct users to doctors, lawyers, or other qualified professionals for specific advice.
Furthermore, the analysis of medical images or conclusions based on such images is now prohibited—to avoid potentially harmful “erroneous diagnoses.”
OpenAI explains that this policy change is related to legal liability and a general aspiration to ensure safer and more ethical use of AI models.
Consequences for Users
Users should no longer expect ChatGPT to tell them which medications to take or how to draft a legal document tailored to their specific situation. Instead, it can assist with clarifying concepts, handling basic informational queries, or preparing supporting materials (for example, brief explanations). This changes ChatGPT’s role from a “conditional assistant-consultant” to an “informational resource without replacing a professional.”
As a reminder, OpenAI recently officially unveiled its own browser, ChatGPT Atlas, which is intended to change the way users interact with the web. The new product, combining artificial intelligence and classic web browsing, is already available for macOS and will soon be released on Windows, iOS, and Android.
