US Accuses Chinese AI Firms of Industrial-Scale Trade Secret Theft
U.S. agencies claim Chinese artificial intelligence companies stole U.S. trade secrets through model distillation, while Beijing rejected the allegations.
U.S. federal agencies have alleged that Chinese artificial intelligence firms acquired American corporate trade secrets on an industrial scale. The Chinese side rejected the allegations, stating that the accusations are completely baseless.
Joint Warning from American Agencies
U.S. federal agencies stated that Chinese artificial intelligence companies have targeted the trade secrets of their American counterparts in order to cut costs and bring technology to market faster.
The FBI, the National Security Agency, and the Cybersecurity and Infrastructure Security Agency announced this situation to the public by issuing a joint warning.
Model Distillation Method and Debates
The published report stated that major Chinese technology companies extracted proprietary functions by training weaker artificial intelligence models on advanced models.
While the model distillation technique can normally be a legitimate research method, U.S. officials argue that the scale of the campaigns by Chinese firms has reached the level of theft.
Gray Market and Transfer Stations
U.S. agencies noted that China-based artificial intelligence firms use a gray market called transfer stations to bypass geographical restrictions and terms of use.
The report emphasized that certain Chinese technology companies, such as DeepSeek and Alibaba Group, have carried out high-volume information distillation campaigns against U.S. artificial intelligence companies.
Sanction Threat and Political Dimension
Treasury Secretary Scott Bessent stated that sanctions and Entity List designations targeting Chinese companies distilling American artificial intelligence models are on the table.
These developments are expected to be addressed during the meetings between U.S. President Donald Trump and Chinese leader Xi Jinping in Washington late this month.
China's Response to the Accusations
A spokesperson for the Chinese Embassy in Washington stated that highlighting the concept of distillation is a deliberate attack on China's development in the artificial intelligence industry.
Meanwhile, Chinese Foreign Ministry spokesperson Mao Ning noted that the country's artificial intelligence development reflects a high degree of self-reliance in science and technology.
Findings from Anthropic and Google
Anthropic claimed that three Chinese artificial intelligence laboratories conducted an industrial-scale campaign to capture the capabilities of its Claude model.
Additionally, Google's Threat Intelligence Group reported that threat actors are increasingly targeting proprietary artificial intelligence models for espionage, extortion, and resource theft.