Brazilian BWC Official Says AI-Assisted Bioweapon Risks Are Real

A Brazilian official involved with the Biological Weapons Convention (BWC) has said the risk of artificial intelligence being used to assist in the creation of biological weapons is real, citing a recent disclosure by US technology company Anthropic.

Frederico Meyer noted that Anthropic had announced it detected and prevented several attempts by users to exploit artificial intelligence systems to assist in creating biological weapons, a disclosure he referenced as evidence that the threat is not merely theoretical.

The comments reflect growing international concern among arms-control and biosecurity officials about the potential for advanced AI models to lower the technical barriers involved in developing dangerous biological agents, an issue that has drawn increasing attention from both AI developers and international security bodies.

The Biological Weapons Convention is an international treaty that prohibits the development, production and stockpiling of biological weapons, and BWC-linked discussions increasingly address emerging technological risks alongside traditional biosecurity concerns.

Meyer's remarks did not detail specific policy proposals to address the risk, though his reference to the Anthropic disclosure suggests officials are looking to real-world examples from AI companies to inform ongoing international discussions about safeguards against AI-assisted bioweapon development.