The White House has accused China of orchestrating a sweeping campaign to steal advanced artificial intelligence technology from American companies, describing the effort as “industrial-scale” and vowing stronger countermeasures. In a memo, Michael Kratsios, director of the White House Office of Science and Technology Policy, said the US government has evidence that foreign actors, primarily based in China, are systematically targeting cutting-edge AI systems developed in the United States.
According to the memo, these operations involve the use of tens of thousands of proxy accounts and sophisticated “jailbreaking” techniques to bypass safeguards and extract proprietary data from leading AI platforms. Officials say the goal is to replicate the capabilities of advanced American models at significantly lower cost, giving Chinese firms a strategic advantage in the global race for AI dominance.
The allegations come amid heightened geopolitical tensions and just weeks before a planned meeting between Donald Trump and Xi Jinping in Beijing. The dispute underscores the growing importance of artificial intelligence as a central battleground in US-China relations, with both nations competing for technological supremacy.
Central to the controversy is the technique known as distillation, where smaller AI models are trained using outputs from more advanced systems. The issue gained prominence following accusations that DeepSeek used this method to build a high-performing model at a fraction of the cost. US officials argue that while distillation has legitimate uses, its unauthorized, large-scale application to extract intellectual property undermines innovation and national security.
American AI firms, including OpenAI and Anthropic, have raised alarms about such practices, warning that they allow foreign competitors to bypass export controls and rapidly close the technological gap. In earlier statements, both companies indicated they had evidence suggesting their models were used without authorization to train competing systems.
Experts say the implications extend beyond economic competition. Chris McGuire of the Council on Foreign Relations warned that distilled models may lack critical safety safeguards, increasing risks related to cyber attacks or even the development of harmful technologies. He suggested the US should consider stricter measures, including sanctions and limiting Chinese access to American AI systems.
China has rejected the accusations, with its embassy in Washington calling the claims “pure slander.” Spokesperson Liu Pengyu stated that China supports technological progress through cooperation and fair competition, and emphasized the country’s commitment to protecting intellectual property rights.
In response to the growing threat, the US government plans to share intelligence with domestic AI firms about ongoing attacks and coordinate defensive strategies. Lawmakers are also taking action, with the House Foreign Affairs Committee advancing legislation aimed at restricting China’s access to sensitive technologies. One proposed measure would expand the use of the US “entity list,” effectively blacklisting companies involved in unauthorized AI distillation.
