Washington, DC, September 9 — US 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 the deployment of AI products.
According to a joint advisory released Tuesday by the Federal Bureau of Investigation, National Security Agency and Cybersecurity and Infrastructure Security Agency, major Chinese technology companies are using “distillation” techniques to train smaller AI models on more advanced systems and extract proprietary capabilities from US frontier AI models.
US authorities and technology executives acknowledge that model distillation can be used for legitimate research, but have argued that the scale and deliberate targeting of Chinese campaigns amount to theft of intellectual property.
The advisory comes amid escalating geopolitical tensions over artificial intelligence. President Donald Trump has indicated that AI will be among the topics discussed when he hosts Chinese President Xi Jinping in Washington, DC, later this month.
US agencies said China-based AI companies are using a gray market of proxies known as “transfer stations” to bypass geographic restrictions, security safeguards and terms of service imposed by American technology companies.
The advisory specifically named Alibaba Group and DeepSeek among Chinese technology companies accused of conducting high-volume knowledge-distillation campaigns against US counterparts.
US Treasury Secretary Scott Bessent warned that punitive measures could follow against Chinese companies involved in such activities. He said sanctions and Entity List designations could be considered if Chinese companies engage in covert, industrial-scale distillation campaigns that cross the line into intellectual property theft.
In a related disclosure, Anthropic alleged that three Chinese AI laboratories had organized an “industrial-scale campaign” to extract capabilities from its Claude AI model.
Anthropic said the Chinese laboratories generated more than 16 million exchanges with Claude through approximately 24,000 fraudulent accounts.
The intensifying global competition for AI leadership has also increased cybersecurity risks for US companies.
A report published Tuesday by Google’s Threat Intelligence Group said hackers with various motivations are increasingly targeting AI assets, including proprietary AI models and source code.
The report said enterprise AI assets, ranging from model weights to cloud-computing quotas, have become high-value targets for espionage, extortion and resource theft. (ANI)
About The Author
