OpenAI updates terms to forbid usage for medical and legal advice (openai.com)

OpenAI updated its Usage Policies (effective in the 2025-10-29 universal rollout) to explicitly bar using its models to provide tailored legal or medical advice that requires a professional license unless a licensed professional is appropriately involved. The change is part of a broader consolidation of policies across products and is enforced through a mix of automated and manual monitoring, developer moderation tools, and appeal processes. OpenAI frames this as a safety-first move to reduce harms, clarify developer responsibilities, and align with professional and legal duties. For the AI/ML community this narrows permissible high-risk applications: builders of chatbots, triage systems, document-review tools, and other integrations must adopt human-in-the-loop workflows, partner with licensed professionals, or limit outputs to general informational content. Practically, expect stricter app review, use of OpenAI’s moderation APIs, and potential access restrictions for services that automate high-stakes decisions (health, legal, employment, finance, etc.). The update signals growing regulatory and liability pressure on model deployment and will likely accelerate certified workflows, provenance controls, and product design patterns that explicitly separate informational assistance from regulated professional services.