The administration of Donald Trump in the United States accused Chinese corporations on the 23rd (local time) of stealing U.S. artificial intelligence (AI) technology.
Michael Kratsios, White House Office of Science and Technology Policy director, said on X (formerly Twitter) that day, "The United States has obtained evidence that foreign firms, primarily China, are conducting large-scale distillation campaigns to steal U.S. AI," adding, "We will take action to protect American innovation."
Earlier, there were reports that leading U.S. AI corporations OpenAI, Anthropic and Google were jointly responding to China's copying of advanced AI models. Bloomberg reported that the three, through the nonprofit Frontier Model Forum, established in 2023 with Microsoft (MS), are sharing information related to China's theft of AI technology.
The information the three corporations are collecting relates to Chinese AI firms' "adversarial distillation," the same as the "large-scale distillation campaign" by Chinese firms that Kratsios pointed out. Distillation is a method of training a new model using the answers of a higher-level AI model as data.
OpenAI identified China's "DeepSeek" as a firm stealing AI technology and even delivered a memo with those details to the U.S. House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party. However, Kratsios did not mention any specific firm by name.
That day, Kratsios said that foreign firms, including China, "are systematically siphoning off America's breakthrough technologies by running organized campaigns using tens of thousands of proxies and jailbreaking—hacking techniques that induce AI systems to bypass security or ethical constraints to carry out malicious purposes."
He added, "Foreign firms founded on such a vulnerable foundation will have little confidence in the integrity and reliability of the models they produce."