US claims Chinese AI firms are carrying out ‘industrial-scale’ theft of trade secrets

Federal agencies in the United States have accused Chinese artificial intelligence firms of conducting large-scale theft of trade secrets from their American counterparts. The FBI, National Security Agency, and Cybersecurity and Infrastructure Security Agency stated that Chinese tech companies use a technique called "distillation" to extract restricted proprietary functionalities and capabilities from US frontier AI models.

Distillation, which involves training smaller AI models on larger, more advanced ones, can be a legitimate research technique, but American officials and AI executives claim that Chinese campaigns against US firms are tantamount to theft. The advisory highlights the sheer scale and sophistication of these campaigns, suggesting they are the core of Chinese AI model development.

China-based AI firms reportedly use "transfer stations" in a gray market to bypass US AI firms' geographic restrictions, safeguards, and terms of use. Treasury Secretary Scott Bessent threatened sanctions against Chinese companies engaged in industrial-scale distillation attacks that cross into IP theft. China's foreign ministry spokesperson, Mao Ning, rejected the accusations, stating that China's AI development reflects its self-reliance and technological strength, and urging the US to refrain from making unfounded accusations against China.

The US-China race for AI supremacy has put American firms at risk of espionage, extortion, and resource theft, according to a report by Google's Threat Intelligence Group.