Overview
- OpenAI’s Oct. 29 usage-policy update bars the provision of tailored advice that requires a license and prohibits automating high‑stakes decisions without human review, including in medicine.
- An OpenAI spokesperson told Business Insider the limits are not a new change and that ChatGPT can still provide general legal and health information.
- News reports describe tighter enforcement that blocks requests for medication dosages, lawsuit templates, or investment tips, with safety filters also curbing hypothetical workarounds.
- Coverage points to documented harms, including a hospitalization after a user replaced table salt with sodium bromide and a case in which a patient delayed care after the bot said cancer was highly unlikely.
- Outlets link the policy line‑drawing to user safety and liability concerns, with Business Insider noting potential constraints on personalized healthcare products.