Washington: United States federal agencies have alleged that Chinese artificial intelligence companies are carrying out 'industrial-scale' theft of American competitors' trade secrets to reduce development costs and accelerate market deployment.
According to a joint advisory released on Tuesday by the Federal Bureau of Investigation, National Security Agency and Cybersecurity and Infrastructure Security Agency, major Chinese technology companies are using 'distillation' techniques, in which smaller models are trained on more advanced architectures, to 'extract restricted proprietary functionalities and capabilities of US frontier AI models'.
While US authorities and technology executives acknowledge that distillation can be used for legitimate research, they have repeatedly argued that the scale and deliberate targeting of Chinese campaigns amount to outright theft.
The advisory adds to these concerns amid escalating geopolitical tensions over artificial intelligence. President Donald Trump has indicated that artificial intelligence will be part of the agenda when he hosts Chinese President Xi Jinping in Washington, DC, later this month.
Describing the operational scale of the campaigns, the advisory stated: "The sheer scale of these campaigns and their sophistication indicate that distillation is not a supplement to these companies' AI model development, but the critical core of it."
US agencies alleged that China-based AI entities use "a grey market of proxies known as 'transfer stations'" to bypass geographical restrictions, defensive safeguards and terms of service imposed by American companies.
The advisory specifically named several Chinese technology corporations, including Alibaba Group and DeepSeek, as carrying out 'high-volume knowledge distillation campaigns' against US counterparts.
On the enforcement front, US Treasury Secretary Scott Bessent warned of possible punitive measures against Chinese companies involved in such activities.
Writing on X, Bessent stated: "When [People's Republic of China] firms conduct covert, industrial-scale distillation attacks that cross the line into IP theft, sanctions and Entity List designations will be on the table."
In a related disclosure, Anthropic claimed that three Chinese AI laboratories had orchestrated an "industrial-scale campaign" aimed at extracting capabilities from its flagship Claude model.
Anthropic alleged that the Chinese laboratories 'generated over 16 million exchanges with Claude through approximately 24,000 fraudulent accounts'.
Meanwhile, intensifying global competition for artificial intelligence dominance has heightened cybersecurity vulnerabilities for US corporations.
A report published on Tuesday by Google's Threat Intelligence Group found that hackers driven by various motives are 'increasingly targeting AI assets'.
The report said malicious actors have increasingly focused on 'proprietary AI models and source code', adding: 'This shift underscores that enterprise AI assets, from model weights to cloud compute quotas, are high-value targets for espionage, extortion, and resource theft'.