Q: On September 8 EST, the U.S. National Security Agency (NSA), Cybersecurity and Infrastructure Security Agency (CISA), and Federal Bureau of Investigation (FBI) issued a joint cybersecurity advisory, warning of China-based AI companies targeting U.S. AI companies with “industrial-scale” knowledge distillation campaigns to acquire the capabilities of U.S. frontier models and proposing defensive measures to U.S. companies. What is China’s comment?

A: We have noted the relevant developments. In China’s view, the U.S. allegation that Chinese AI companies are engaged in “industrial-scale” distillation campaigns targeting U.S. models is groundless and legally unsound. This move politicizes and weaponizes distillation, a normal technical and commercial issue in the industry, and applies double standards in practice. The U.S. advisory is yet another proof that the country is pursuing technological hegemony, seeking a computing power monopoly and suppressing competition in the AI field. China firmly opposes this.

Distillation is a common practice in the AI sector whereby models learn from one another. It is intrinsically a neutral technical means, and model companies around the world, including U.S. businesses, are using it. Such techniques improve learning efficiency make the use of human knowledge more effective, and is of positive significance for growing AI industries in all countries, unlocking the potential of AI technologies, closing development gaps and more broadly benefiting developing countries and the general public.

What the U.S. has done is a typical case of applying double standards. On AI, China encourages open source, openness, cooperation and sharing. China’s open-source models are available to global companies, including U.S. companies, and facilitate the development of related models. Reports from U.S. companies on their own model development also disclose their extensive distillation of Chinese models. By contrast, the U.S. has repeatedly and unilaterally accused Chinese companies of “industrial-scale” distillation campaigns, smearing an industry-wide practice as an attack. This reflects both the anxiety of the U.S. side and its double standards.

What the U.S. has done is a typical case of the U.S. pursuing industrial monopoly under the pretext of fighting against distillation. By having security agencies issue the advisory, the U.S. ties the interests of individual companies and capital to national security and mobilizes state power to interfere in normal commercial activities. We note that certain U.S. AI companies abuse their position of competitiveness and impose overbroad geographic restrictions and other unfair terms in user agreements. The U.S. distillation allegations could be interpreted as endorsing those unfair terms.

What the U.S. has done is a typical case of mobilizing state power to defend technological hegemony and computing power monopoly and suppress competition. Since the beginning of this year, U.S. executive and legislative branches have repeatedly threatened sanctions and suppression against Chinese companies for distillation. The aim is to crush competition and maintain U.S. monopoly over computing power and data, allowing only the U.S. itself to win while denying humanity shared gains. This turns global AI resources that should have been openly and fairly accessible into exclusive assets of a few companies and blocks other countries from the universal benefits of AI innovations.

The large language model industry is now at a critical stage of the combined powers of technological iteration, recurring risks and continued growth. As major countries, China and the U.S. should address the challenges and risks in the AI sector in a positive and responsible manner. The Chinese and U.S. heads of state have agreed to launch an intergovernmental AI dialogue. China stands ready to engage in constructive and professional discussions with the U.S. on the basis of equality, mutual benefit and win-win cooperation. But if the U.S. side takes actions to contain and suppress Chinese AI companies under the pretext of combating distillation, China will resolutely take countermeasures. We hope the U.S. side will move in the same direction as the Chinese side, manage risks through dialogue and communication and let the AI industry benefit the whole world.(Released on September 9)