The United States accuses Chinese AI companies of maliciously plagiarizing advanced American model technology

According to a joint statement issued by three U.S. security agencies, U.S. cyber and law enforcement officials today accused Chinese artificial intelligence (AI) companies of using "distillation" methods to maliciously plagiarize advanced U.S. AI model technology.

Reuters reported that "distillation" refers to using the results produced by high-cost large-scale AI models to train smaller AI models to reduce the training costs of developing new AI tools.

The U.S. National Security Agency (NSA), Cybersecurity and Infrastructure Security Agency (CISA), and the U.S. Federal Bureau of Investigation (FBI) accused Chinese companies such as DeepSeek, Moonshot AI, Alibaba, MiniMax, and StepFun of using various variants of U.S. AI models to train their own AI products.

These agencies said that the above-mentioned Chinese companies "likely distilled American AI models with the knowledge of the Chinese government."

"Chinese AI companies are conducting aggressive, malicious and targeted distillation activities on a large scale," officials said.

The US accusation comes as US President Trump's administration is preparing for Chinese President Xi Jinping's expected visit to the United States on September 24. Reuters previously disclosed that the United States and China are expected to hold a dialogue on AI security risks in mid-September. (Compiled by: Hong Qiyuan) 1150909