US alleges Chinese AI firms stole American technology at industrial scale
US officials accuse six Chinese AI firms, including DeepSeek and Alibaba, of copying American AI models via distillation, ahead of Xi's US visit.
Washington has formally accused six Chinese artificial intelligence companies of engaging in the large-scale theft of American AI technology, a move that is likely to cast a shadow over the upcoming meeting between the US and Chinese presidents later this month.
In a joint statement, US law enforcement and intelligence officials alleged that the Chinese firms — including DeepSeek, Moonshot AI, and Alibaba — used a technique known as 'distillation' to copy intellectual property from leading American AI labs. Distillation involves training a smaller, cheaper AI model using the outputs of a larger, more expensive one, thereby drastically cutting the cost and time needed to develop a new system.
The officials said the companies targeted models built by Anthropic, OpenAI, Google, and SpaceX, and that the activity was carried out 'likely with Chinese government awareness.' The statement described the practices as 'aggressive, malicious and targeted distillation activities at an industrial scale.'
The allegations come at a delicate diplomatic moment. Chinese President Xi Jinping is scheduled to visit the US in late September, and the two countries are also preparing for a mid-September dialogue on AI safety risks. Washington had previously levelled a similar accusation against Beijing in April, ahead of President Donald Trump's visit to China.
Beyond the commercial implications, US officials warned that the copied technology could bolster China's military and cyberattack capabilities, potentially posing a threat to the US and its allies. This concern is supported by earlier reporting that Chinese military researchers have used outputs from leading US AI models to advance domestic defence systems.
The Chinese embassy in Washington has not yet responded to requests for comment on the allegations.