US frontier AI companies warn authorities over sophisticated distillation attacks — China warns of 'countermeasures' if America tries to constrain domestic AI models
TL;DR
The NSA, CISA and FBI have issued a joint advisory warning that China-based AI companies are running industrial-scale distillation campaigns against US frontier models. According to the September 8 advisory, DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z. AI extracted billions of tokens from models including Claude, GPT, Gemini and Grok to train cheaper alternatives. China's Ministry of Commerce rejected the claims and threatened countermeasures if the US uses them to justify new restrictions.
Nauti's Take
One step forward: authorities now name specific labs and techniques, which helps providers detect API abuse more precisely. The risk is escalation, since export rules from Washington and countermeasures from Beijing would also hit companies that use cheaper models like DeepSeek or Qwen legitimately.
Teams relying on Chinese open-weight models should keep an alternative in their stack.
Summary
The NSA, CISA and FBI have issued a joint advisory warning that China-based AI companies are running industrial-scale distillation campaigns against US frontier models. According to the September 8 advisory, DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z.
AI extracted billions of tokens from models including Claude, GPT, Gemini and Grok to train cheaper alternatives. China's Ministry of Commerce rejected the claims and threatened countermeasures if the US uses them to justify new restrictions.
Sources