US Intelligence Agencies Issue Distillation Warning Against Chinese AI Models
Three national security and intelligence agencies announced that Chinese AI firms are employing data extraction and distillation tactics against American models.
Three federal US national security and intelligence agencies have issued a joint warning that China-based artificial intelligence companies are conducting aggressive distillation campaigns against American models.
Joint Warning and Targeted Companies
The National Security Agency, the Cybersecurity and Infrastructure Security Agency, and the Federal Bureau of Investigation issued a joint bulletin.
The bulletin stated that Chinese companies such as DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI were targeted.
Distillation Tactics and Methods
It was stated that Chinese companies use malicious distillation tactics to extract billions of tokens from US models.
It was reported that these operations are concealed through local APIs, cloud providers, and third-party aggregators.
Targeted American Models
It was noted that leading models such as Anthropic Claude, OpenAI ChatGPT, Google Gemini, and SpaceXAI Grok were targeted.
It was emphasized that the intellectual property rights of these systems are under threat.
Recommended Measures and Precautions
It was recommended to establish comprehensive detection mechanisms for American AI developers.
It was requested that inter-agency intelligence sharing and targeted response modifications be implemented.