China Rejects US Allegations of 'Industrial-Scale' AI Model Distillation

China has rejected claims by United States security agencies accusing Chinese artificial intelligence developers of conducting 'industrial-scale' distillation of American frontier models. In a formal response on September 9, China's Ministry of Commerce characterized knowledge distillation as a globally recognized, neutral engineering practice utilized across the AI sector, including by US firms. Beijing accused Washington of attempting to protect a computing monopoly under the pretext of national security, while Chinese foreign affairs officials maintained that domestic AI progress is rooted in scientific self-reliance.

The confrontation signals an expansion of US-China tech frictions beyond hardware sanctions like advanced semiconductor export controls into software development workflows, foundational model architectures, and intellectual property norms. Labeling model distillation a national security threat could pave the way for US restrictions on cross-border API access or terms-of-service enforcement, complicating global research cooperation and fracturing software development pipelines.

Beijing has strongly pushed back against allegations from Washington that Chinese artificial intelligence enterprises are systematically extracting capabilities from American frontier models through large-scale knowledge distillation. The response followed a joint cybersecurity advisory issued on September 8 by the US National Security Agency, Cybersecurity and Infrastructure Security Agency, and Federal Bureau of Investigation. The US agencies accused Chinese AI firms of engaging in 'industrial-scale' distillation to mirror US capabilities and advised American companies to adopt defensive measures against such practices. Addressing the advisory on September 9, a spokesperson for China's Ministry of Commerce stated that the US accusations are unfounded in fact and law. The ministry described model distillation as a neutral, standard industry technique used globally—including by US enterprises—to improve training efficiency, accelerate model learning, and utilize collective knowledge. Beijing argued that the technique plays a constructive role in narrowing technological divides and broadening access to AI innovation worldwide. The commerce ministry criticized Washington for exercising what it called a double standard. The spokesperson noted that China encourages open-source collaboration, offering models to international developers, whereas US firms have themselves disclosed in research reports that they distill Chinese models. According to the ministry, Washington's actions link private corporate interests to state power, effectively endorsing restrictive licensing agreements and broad territorial barriers imposed by dominant US platforms. Chinese officials warned that this pressure reflects an effort to preserve monopoly control over compute and datasets rather than legitimate security governance. The commerce ministry warned that if the US uses accusations over distillation to enact punitive measures or suppress Chinese AI enterprises, Beijing will resolutely take countermeasures to protect domestic firms. Simultaneously, China's Ministry of Foreign Affairs emphasized that the country's advancements in artificial intelligence are the direct result of high-level scientific and technological self-reliance, rather than external extraction. Both agencies reaffirmed Beijing's openness to constructive discussions through the bilateral intergovernmental AI dialogue agreed upon by the leaders of both nations, urging Washington to manage differences through mutual consultation.