The United States government has announced decisive measures to prevent foreign technology companies, particularly those from China, from exploiting artificial intelligence models developed in the country. This move comes amidst a growing rivalry as China closes in on the US in the AI sector.
US Government’s Stance on AI Exploitation
In a recent memorandum, Michael Kratsios, the chief science and technology adviser to the President, accused foreign entities, mainly from China, of systematically extracting capabilities from leading US AI systems. The administration plans to collaborate with American AI firms to detect such activities, bolster defenses, and implement punitive measures against offenders.
This announcement coincides with a period where China’s advancements in AI are challenging US dominance. The White House emphasizes the importance of maintaining a lead in AI to set global standards and secure economic and military advantages. A study by Stanford University’s Institute for Human-Centered AI highlights that the gap in performance between US and Chinese top AI models is nearly closed.
China’s Response to US Allegations
China has strongly opposed the US’s actions, which it perceives as unjust suppression of Chinese companies. Liu Pengyu, a spokesperson for the Chinese embassy in Washington, reiterated China’s commitment to fostering scientific progress through cooperation and emphasized the importance of intellectual property protection.
Guo Jiakun, a spokesperson for China’s Foreign Ministry, dismissed the US claims as baseless and accused the US of tarnishing China’s achievements in AI. He urged the US to abandon its prejudices and promote scientific exchange between the two nations.
Legislative and Industry Reactions
The memo was released during the same week that the House Foreign Affairs Committee unanimously supported a bill aiming to identify and sanction foreign actors involved in extracting technical features from US-owned AI models. Representative Bill Huizenga highlighted the threat posed by model extraction attacks, calling it a form of economic coercion and intellectual property theft by China.
Last year, Chinese startup DeepSeek alarmed US markets by introducing a large language model rivaling US AI giants at a lower cost. Allegations arose that DeepSeek and other Chinese labs were illicitly extracting capabilities from US models.
In a counterpoint, US-based Anysphere admitted to using an open-source model from Chinese company Moonshot AI for its latest product, showing the complex nature of AI technology exchange.
Future Outlook and Challenges
Kyle Chan, a technology expert at The Brookings Institution, noted the difficulty of distinguishing unauthorized distillation from legitimate data requests, likening it to finding needles in a haystack. However, he suggested that information sharing among US AI labs and federal involvement could mitigate these challenges.
The effectiveness of the proposed US legislative measures remains uncertain, especially with President Trump’s anticipated visit to China. The dynamics of US-China relations in AI innovation continue to evolve, with both countries vying for technological supremacy.
