The White House's export restrictions on Anthropic's Mythos were partly driven by fears of access by a China-linked group. If the Chinese government accessed Mythos, it would pose a serious national security risk.
The White House's decision to impose export restrictions on Anthropic's Mythos was partly driven by fears that a China-linked group had accessed the model. According to a Semafor report, if the Chinese government accessed Mythos 5 or Fable 5, it would pose a serious national security risk, potentially through reverse engineering via distillation.
The White House has not confirmed the report, and Trump adviser David Sacks focused on jailbreak concerns rather than China. Anthropic stated that China was not mentioned during discussions on export controls. This is not the first security breach of Anthropic's powerful AI models.
If confirmed, this incident would represent a significant challenge to US AI technological superiority. It underscores the importance of AI model security and export control policies.