- US officials initially blocked Anthropic's advanced AI models, Fable and Mythos, citing national security risks due to their potential to identify critical infrastructure vulnerabilities.
- An export ban was imposed, which Anthropic found difficult to enforce given the inability to verify user nationalities.
- The ban has since been lifted after Anthropic collaborated with the US Commerce Department, implementing enhanced safeguards and controls to prevent misuse.
- Mythos 5 is now available to "trusted" US organizations, with plans for broader access, while Fable 5, designed for the general public with stronger protections, will be released soon.
- Anthropic is deepening its collaboration with the US government and other tech companies to develop shared standards for assessing and fixing AI vulnerabilities, acknowledging that fully robust AI models are challenging to achieve.
IN FULL
