Brazil’s medical council brings national AI-in-medicine rules into force

Brazil’s new medical AI resolution is one of the clearest national attempts to spell out how generative and diagnostic systems can and cannot be used in clinical practice. By explicitly treating AI as a tool subordinate to physician judgment and tying misuse to professional ethics sanctions, the Federal Medical Council is asserting that responsibility stays with humans, even as hospitals lean on chatbots for transcription, summarization and even triage.

For model providers like OpenAI, Anthropic and Perplexity, this is an early preview of how sector specific rules will intersect with general purpose models. The resolution reaches beyond bespoke medical AI to everyday uses like drafting clinical notes or summarizing history, and it does so without a grace period for institutional governance. That will push Brazilian providers to document how they select, validate and monitor any AI system in the workflow, including off the shelf chatbots.

In AGI terms, this is not a brake on research, but it does sketch a world where highly capable systems are boxed in by domain regulators who care more about accountability than innovation. If other countries follow Brazil’s lead, frontier labs will face a patchwork of sector rules on top of horizontal AI acts. That could slow deployment of powerful general models into safety critical fields even as capabilities continue to improve.