According to information from the Ministry of Commerce's website on September 9, a spokesperson for the Ministry of Commerce answered reporters' questions regarding the cybersecurity announcement issued by the United States concerning Chinese artificial intelligence companies’ activities in the US.
Question: On September 8th, Eastern US time, the US National Security Agency, the Cybersecurity and Infrastructure Security Agency, and the Federal Bureau of Investigation jointly issued a statement, accusing Chinese artificial intelligence companies of conducting "industrial-scale" distillation activities in the US, acquiring advanced US models, and offering suggestions on defense measures to US companies. What is China's comment on this?
Answer: We have taken note of the relevant situation. China believes that the US accusations that Chinese artificial intelligence companies are engaged in "industrial-scale" model distillation are baseless and legally unfounded. The US actions involve politicizing and tooling this normal technical and commercial issue within the industry, and applying double standards in practice. The relevant announcements by the US are further evidence of their efforts to promote technological hegemony in the field of artificial intelligence, to monopolize computing power, and to suppress competition. China firmly opposes such practices.
Distillation is a common practice for models in the field of artificial intelligence to learn from each other. It is essentially a neutral technical approach, and global model companies, including those from the United States, use this technique. This approach can help models improve their learning efficiency and enable more efficient utilization of human knowledge. It is of positive significance for countries to develop the artificial intelligence industry, unlock the potential of artificial intelligence technology, bridge development gaps, and benefit developing countries and the general public more widely.
The American approach is a typical example of double standards. Regarding artificial intelligence, China encourages open source, cooperation, and sharing. Chinese open-source models are made available to global companies, including those from the United States, thereby facilitating the development of related models. Reports on the development of relevant models by American companies also reveal that they frequently distill Chinese models. In contrast, the Americans repeatedly single out Chinese companies for engaging in “industrial-scale” distillation, distorting common industry practices as offensive actions. This not only reflects the anxiety of the Americans but also demonstrates their use of double standards.
The American approach is a typical example of using the pretext of combating distillation to achieve industrial monopoly. The US security department’s announcements tie the interests of individual companies and capital with national security, thereby leveraging state-authoritative institutions to interfere with normal commercial activities. We note that some American artificial intelligence companies abuse their competitive advantages by establishing arbitrary and oppressive terms in user agreements that impose broad geographical restrictions. The accusations made by the US regarding distillation seem to endorse these oppressive provisions.
The American approach is a typical example of using national power to maintain technological dominance and monopolies in computing power, thereby suppressing competition. Since the beginning of this year, the U.S. executive and legislative bodies have frequently issued sanctions and threats against Chinese enterprises for their “distillation” activities. The goal is to suppress competition and maintain the U.S.'s monopoly in computing power and data. Only the United States can win, not humanity. This approach turns the global AI resources that should be openly and fairly available into exclusive resources for a few companies, preventing other countries from participating in the sharing of AI innovation achievements.
Currently, the large-model industry is in a critical period of technological iteration, frequent risks, and growth development. As major nations, China and the US should approach the challenges and risks faced by artificial intelligence with a proactive and responsible attitude. The heads of state of both countries have agreed to initiate inter-governmental dialogues on artificial intelligence. The Chinese side is willing to engage in constructive and professional discussions with the US on the basis of equality, mutual benefit, and cooperation. However, if the US takes actions to suppress Chinese artificial intelligence companies under the pretext of targeting certain sectors, the Chinese side will definitely take measures to counter such actions. It is hoped that the US and China will move towards a common goal through dialogue and communication, thereby managing risks and promoting the development of the artificial intelligence industry for the benefit of the entire world.
