The Verge
White House restricted Anthropic's Mythos due to suspected Chinese access.
- Security breach: Export restrictions on Anthropic's Mythos were triggered by fears that China-linked groups had accessed the advanced AI model.
- National security risk: Chinese access to Mythos 5 or Fable 5 would represent a serious threat to U.S. interests.
- Reverse engineering: The government could use distillation techniques to train 'student' AIs that replicate the advanced model's capabilities.
For ethics
Teams should audit which AI models they're using internally and whether they have government security clearances, as export restrictions could disrupt workflows.