US Intelligence Agencies Issue Warning Over Industrial-Scale AI Model Distillation by Chinese Firms

In a significant joint cybersecurity advisory, the U.S. National Security Agency (NSA), Cybersecurity and Infrastructure Security Agency (CISA), and Federal Bureau of Investigation (FBI) have formally accused six China-based artificial intelligence companies of engaging in aggressive, industrial-scale "knowledge distillation" campaigns against leading American AI models. The advisory specifically names prominent Chinese developers, including DeepSeek, Alibaba, Moonshot AI, MiniMax, StepFun, and Z.AI, alleging that they have extracted billions of tokens through millions of targeted requests from U.S. frontier AI models such as OpenAI's GPT, Anthropic's Claude, Google's Gemini, and xAI's Grok since late 2024. Knowledge distillation is a standard machine learning practice where a smaller "student" model learns from a larger, more capable "teacher" model. However, U.S. intelligence agencies assert that Chinese firms have operationalized this technique to shortcut massive research and development expenditures, compute costs, and energy requirements. The advisory highlights that these companies routed requests through proxy networks, fraudulent accounts, and third-party aggregators to bypass terms of service and security controls. In response, U.S. officials recommended that American AI firms deploy behavioral monitoring and subtle response degradation to detect and counter unauthorized distillation activities. In turn, China's Ministry of Commerce rejected the allegations, defending distillation as a standard industry method and accusing Washington of protectionism.