US accuses Chinese AI firms of ’malicious’ copying of AI technology

U.S. authorities have accused six Chinese artificial intelligence firms of maliciously replicating technology from American AI companies, potentially straining ties between the world's two largest economies ahead of an upcoming summit. Distillation, a method of training smaller AI models using larger, more expensive ones as a cost-cutting measure, is alleged to be the tool used by the Chinese companies.

Six firms, including DeepSeek, Moonshot AI and Alibaba, are named in the U.S. statement as engaging in this practice, likely with the knowledge and backing of the Chinese government. The companies targeted AI models from major American firms such as Anthropic, OpenAI, Google and SpaceX. The U.S. officials described these actions as aggressive and at an industrial scale, carried out by Chinese-based AI companies.

The allegations come as U.S. President Donald Trump's administration prepares for Chinese President Xi Jinping's visit to the U.S. in late September, and as discussions on AI safety risks are scheduled for mid-September. The Chinese embassy in Washington has not yet commented on the allegations. The U.S. previously made a similar claim in April, ahead of Trump's visit to Beijing.

The Tuesday statement suggested that the distillation process not only cuts research and development costs for Chinese AI firms but also boosts China's military and cyberattack capabilities, potentially posing a threat to the U.S. and its allies.