White House accuses China of copying American AI models in ‘industrial-scale’ campaign

The White House has raised serious concerns over what it describes as an 'industrial-scale' campaign by Chinese entities to replicate advanced AI models developed by American companies. Michael Kratsios, the director of the Office of Science and Technology Policy, detailed these allegations in a memo released on Thursday. According to Kratsios, the tactics employed involve the use of tens of thousands of proxy accounts designed to evade detection, alongside sophisticated tools aimed at uncovering proprietary information. He emphasized that these coordinated efforts are extracting capabilities from US AI models, thereby exploiting American innovation and expertise. The competition between the United States and China in the realm of AI technology has intensified, with both nations vying to dominate this burgeoning field. This rivalry has further complicated trade relations, drawing in major tech firms like Nvidia into the fray. Central to the allegations is a method known as 'distillation,' which allows for the transfer of knowledge from a larger AI model to a smaller, cost-effective one. While this technique is common in AI training, companies like Anthropic and OpenAI have previously accused others of using it to unfairly replicate their models’ functionalities. DeepSeek, a Chinese AI startup that previously made headlines on Wall Street, has faced scrutiny for these practices. In February, OpenAI alerted lawmakers that DeepSeek was attempting to mirror the performance of its models, as well as those from other leading US AI firms, through distillation. Anthropic also reported identifying similar large-scale efforts by DeepSeek and two other labs to illicitly extract capabilities from its AI model, Claude. Concerns regarding the security of distilled AI models have also been raised, as these versions may lack the protective measures of the original models. In his memo, Kratsios warned that these campaigns could lead to the removal of essential security protocols and mechanisms designed to ensure that AI models remain ideologically neutral and dedicated to truth-seeking. In response to these developments, the Trump administration is planning several actions aimed at addressing the issue. These include enhancing information sharing with US AI firms about unauthorized distillation attempts, improving collaboration among private sector entities, and exploring ways to hold foreign actors accountable for such actions. The administration’s focus on solidifying the US position as a leader in AI has been a pivotal aspect of its agenda, with calls for federal regulations to streamline innovation while facing criticisms over potential accountability evasion by AI companies. Additionally, there are ongoing controls over AI chip sales to China.