중국, 미국의 AI 기업 ‘증류’ 단속에 “기술패권 위한 탄압, 강력 맞대응”

The United States has accused China of targeting "technological dominance and computational power monopolies" to maintain its competitive edge in artificial intelligence (AI), according to a Chinese statement. As the U.S.-China summit approaches, the issue of AI "concentrated" has emerged as a new point of contention in the two countries' technological rivalry.

The Ministry of Industry and Information Technology released a statement on September 9, asserting that the claim by the U.S. that Chinese AI companies have "industrially scaled" them is both factually and legally unfounded. The ministry warned that if the U.S. uses the pretext of concentration to suppress and pressure Chinese AI companies, it will respond firmly.

The National Security Agency (NSA), Federal Bureau of Investigation (FBI), and Cybersecurity and Infrastructure Security Agency (CISA) jointly released a cybersecurity guideline on August 8, identifying six Chinese companies—DeepSeek, MarketMergeAI, Alibaba, MinMax, StepFun, and ZhiPeng—that engaged in "concentrated" activities against U.S. AI companies by circumventing regional restrictions and usage agreements to extract tens of billions of tokens.

The U.S. claims that this is a "malicious activity" distinct from the general practice of "concentration" in AI. The ministry maintains that concentration is a neutral technical means in AI, widely used by models in the field, including those owned by U.S. companies. They also pointed out that Chinese open-source models are available to companies worldwide, and that Chinese models have been extensively concentrated in U.S. companies' research and development reports.

The ministry accused the U.S. of politicizing and weaponizing normal technological and commercial issues. The guideline highlights that the development of Chinese AI models could enhance China's military capabilities, making AI competition "of national security implications." China labels this as "another clear evidence of seeking scientific and technological hegemony and monopolizing computing power."

The U.S. security agencies' direct engagement further indicates their intent to involve specific companies and capital in national security, the ministry argues. Despite the tension, the ministry noted that both sides have expressed willingness to engage in constructive and professional discussions based on principles of equality, mutual benefit, cooperation, and common benefit.

The bilateral dialogue on AI safety, scheduled for mid-February, will likely focus on distinguishing between the normal and malicious uses of "concentration" in the AI industry.