On September 9, a spokesperson for the Ministry of Commerce responded to questions from reporters regarding a cybersecurity announcement by the U.S. concerning distillation activities of Chinese AI companies. The spokesperson stated that distillation is a common practice in the AI field where various models learn from each other, and it is essentially a neutral technical means used by global model companies, including those in the U.S. This technique can help models improve learning efficiency, enable more effective use of human knowledge, and has positive implications for countries developing their AI industries, unlocking the potential of AI technology, bridging development gaps, and benefiting developing countries and the general public. The actions of the U.S. reflect a typical double standard. China encourages open-source, collaborative sharing in AI, and Chinese open-source models are accessible to global companies, including those in the U.S., facilitating relevant model research and development. Reports from U.S. companies also reveal that they extensively distill Chinese models. In contrast, the U.S. has repeatedly unilaterally accused Chinese companies of engaging in 'industrial-scale' distillation, tarnishing a common industry practice as an aggressive act, which not only reflects U.S. anxiety but also exemplifies double standards.
All Comments