The White House's decision to restrict exports of Anthropic's AI model "Mythos" is reported to have reflected, in part, concern that a China-linked group may have gained access to Mythos. The Verge reported this on June 14, citing a Semafor report.
If the Chinese government actually accessed "Mythos 5" or "Fable 5," it could pose a serious national security risk.
The concern focuses less on access itself than on potential use afterward. If the Chinese government gained access to the model, it could try reverse engineering through a distillation method that trains a "student" AI to mimic more advanced model behavior.
The White House did not officially confirm this. David Sacks (데이비드 삭스), an adviser to Donald Trump's camp, also did not mention China in a post on social media X (Twitter). Instead, he referred to reported issues that "Fable" and "Mythos" could be jailbroken, but Anthropic denied this.
An Anthropic spokesperson said the U.S. government did not raise the China issue during discussions on export controls.
.
Chi-gyu Hwang (황치규)
delight@d-today.co.kr