Trump drops restrictions on Anthropic’s Mythos and Fable models

The US government has dropped its requirement for Anthropic to obtain a license before exporting its Mythos and Fable AI models abroad. This change allows Anthropic to once again make these models publicly available, which it had ceased doing after the US added them to its list of export-restricted technologies on June 12. As part of the agreement, Anthropic has committed to proactively detecting and addressing security risks associated with these models, collaborating with the US government on protocols and standards, and reporting any malicious activity. Specifically, Secretary of Commerce Howard Lutnick stated that Anthropic agreed to work diligently with the US government on protocols and standards for Mythos, Fable, and future models.

The decision to ease restrictions on Anthropic's AI models reflects the competitive landscape of the global AI industry. With Asian AI companies, such as those releasing Fugu and Tulonfeng models, approaching Mythos-level capabilities, the US government faced pressure to ensure American AI companies could compete globally. This move follows the Trump administration's approval of OpenAI's latest models for release to select organizations. The administration's approach to AI policymaking has been erratic, as highlighted by an executive order issued in June that signaled a desire to review models ahead of release. Influential analysts like Dean W. Ball, who recently started a policy position at OpenAI, have criticized this approach.

The lifting of export restrictions on Mythos and Fable reduces uncertainty for Anthropic and potentially sets a precedent for how the US government will handle similar cases in the future. However, it also raises questions about the balance between promoting American AI competitiveness and addressing national security concerns. As Anthropic begins to restore access to its models, it will be important to monitor how the company implements its commitments to security and collaboration with the US government. Additionally, the ongoing development and release of advanced AI models by various companies, including Anthropic, OpenAI, and Asian AI companies, will likely continue to shape the global AI landscape.