FILE – White House Director of Science and Technology Policy Michael Kratsios speaks during a meeting of the White House Task Force on Artificial Intelligence Education in the East Room of the White House on September 4, 2025 in Washington.
Alex Brandon/AP
hide caption
toggle caption
Alex Brandon/AP
WASHINGTON — The Trump administration is vowing to crack down on the exploitation of American artificial intelligence models by foreign tech companies, and singling out China at a time when the country is closing the gap with the US in the AI race.
In a memo Thursday, Michael Kratsios, the president’s chief science and technology adviser, accused foreign entities “primarily based in China” of engaging in deliberate, industrial-scale campaigns to “distill” leading American-made AI systems or extract capabilities and “exploit American expertise and innovation.”
The administration will work with U.S. AI companies to identify such activities, build defenses, and find ways to punish offenders, Kratsios wrote.
The memorandum comes at a time when China is challenging US dominance in artificial intelligence, an area where the White House says the US must prevail to set global standards and gain economic and military advantages. But according to a recent report from Stanford University’s Institute for Human-Centered AI, the US-China gap in the performance of top AI models has “effectively closed.”
China’s embassy in Washington said it opposes “unfair suppression of Chinese companies by the US”.
Embassy spokesman Liu Pengyu said, “China has always been committed to promoting scientific and technological progress through cooperation and healthy competition. China attaches great importance to the protection of intellectual property rights.”
Kratios’ memo also came the same week the House Foreign Affairs Committee offered unanimous, bipartisan support for a bill to establish a process to identify foreign actors who extract “key technical features” of closed-source, U.S.-owned AI models and punish them with measures including sanctions.
“Model extraction attacks are the latest frontier of Chinese economic coercion and theft of American intellectual property,” said Rep. Bill Huizenga, R-Mich., who sponsored the bill. “American AI models are demonstrating transformative cyber capabilities, and it is critical that we prevent China from stealing these technological advancements.”
Last year, Chinese start-up DeepSeek caused a stir in US markets when it released a massive language model that could compete with US AI giants but at a fraction of the cost.

David Sachs, who at the time served as an AI and crypto advisor to President Donald Trump, suggested that DeepSeek copied the American model. Sachs then said, “There is ample evidence that what DeepSeek did here was distill knowledge from OpenAI’s models.”
In a letter to US lawmakers in February, OpenAI, the developer of ChatGPIT, made similar allegations and said China should not be allowed to pursue “autocratic AI” by “appropriating and re-packaging American innovation”.
Cloud chatbot maker Anthropic in February accused DeepSeek and two other China-based AI labs of engaging in a campaign to “illegally extract the capabilities of the cloud to improve their own models” by using distillation technology, which involves “training a less capable model on the output of a stronger model.”
Anthropic said that distillation can be a legitimate way to train AI systems, but it is a problem when competitors “use it to obtain powerful capabilities from other labs in a fraction of the time, and at a fraction of the cost, that it would take to develop them independently.”
But it can go either way. San Francisco-based startup Anisphere, maker of the popular coding tool Cursor, recently admitted that its latest product was based on an open-source model created by Chinese company Moonshot AI, creator of the chatbot Kimi.
Kyle Chan, a fellow at The Brookings Institution, a Washington-based think tank, and an expert on China’s technology developments, said separating unauthorized distillations from the vast volume of legitimate requests for data would be like “looking for needles in a giant haystack.” But information sharing and coordination among U.S. AI labs could help, and the federal government could play an important role in facilitating anti-distillation efforts across labs, Chan said.
It’s hard to assess how far the House bill might go, but Chan said Trump wouldn’t want to swing by a boat with Chinese President Xi Jinping ahead of a planned state visit to Beijing in mid-May.
