U.S. federal agencies have accused Chinese AI companies of systematically copying AI models from American firms.

The Cybersecurity and Infrastructure Security Agency (CISA), the National Security Agency (NSA), and the FBI issued a joint security alert on Tuesday, warning that Chinese AI firms have been targeting U.S. companies to "systematically extract proprietary models through knowledge distillation campaigns."

Knowledge distillation is a machine learning technique involving training a smaller AI model using output from more capable ones, according to an accompanying statement.

"While it's a valid training method, it can be misused to attempt to acquire capabilities from competitors in less time and with less cost than developing them legitimately," the agencies said. "The malicious campaigns described in this advisory form the core of China’s AI development strategy."

The American officials specifically named six companies - DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI - alleging that since at least late 2024, they have "extracted billions of tokens" across millions of exchanges and requests from U.S. models, including Claude, GPT, Gemini, and Grok, "likely with Chinese government awareness."

The agencies said that such tactics violated the U.S. companies' terms of use and undermined fair competition.

"We strongly urge AI companies to take immediate steps to safeguard their platforms against knowledge distillation campaigns that threaten to close the gap in advancements made by American companies," CISA Acting Director Nick Andersen said in the statement.

The U.S. agencies also alleged that the distillation campaign could enhance China's "military and cyberattack capabilities that could be used against the U.S. and our allies." In July, Reuters reported that Chinese military researchers have used outputs from OpenAI's and Anthropic's AI models to train domestic AI systems.

In response to the U.S. allegations, Chinese Foreign Ministry spokesperson Mao Ning said at a Wednesday press briefing that China's AI development is the result of its pursuit of high-level technological self-reliance.

"We hope the U.S. will earnestly implement the important consensus reached by the leaders of the two countries and refrain from making false accusations against and smearing China," said Mao.