The White House has accused China of engaging in “industrial-scale” theft of artificial intelligence technology, alleging that Chinese entities are employing methods such as distillation, proxy accounts, and jailbreaking techniques to extract proprietary knowledge from U.S. frontier labs. This accusation aligns with a recent shift in U.S. policy under the Trump administration that emphasizes intelligence sharing with leading AI firms like OpenAI, Anthropic, and Google to counter such foreign distillation efforts. Additionally, congressional hearings have brought attention to China’s strategies for acquiring U.S. AI advancements, highlighting the challenges posed when direct competition is not feasible.
China: The People’s Republic of China leads global AI development through state-backed labs and firms aggressively advancing frontier technologies. US officials identify Chinese entities as spearheading coordinated distillation operations against American AI labs, leveraging proxies and jailbreaks to extract proprietary know-how for cheaper model replicas. This claim escalates tensions amid ongoing US-China tech rivalry and precedes a planned Trump-Xi summit.
White House: The White House serves as the executive office of the President of the United States, Donald Trump, coordinating national policies on science, technology, and security. Recently, its Office of Science and Technology Policy issued a memo accusing foreign entities primarily in China of running deliberate, industrial-scale campaigns to distill capabilities from US frontier AI models using proxy accounts and jailbreaking techniques. The administration pledges to share intelligence with American AI companies and explore accountability measures to counter these efforts.
US Policy Shift: The Trump administration commits to intelligence sharing with frontier AI firms like OpenAI, Anthropic, and Google to combat foreign distillation campaigns.
Congressional Focus: Recent House Select Committee hearings highlighted China’s tactics to acquire US AI advantages when unable to compete openly.
Distillation Method: Distillation exploits AI model outputs by querying systems at scale to train imitation models without accessing weights, using proxies to evade detection and jailbreaks to bypass safeguards.