As two leading global powers in artificial intelligence, China and the United States cannot afford to cease dialogue despite existing competition. Although disagreements recently arose between them regarding model distillation, this dispute is not the central issue in their AI relationship.
Distillation is a widely recognized, universal methodology that Gai Keke from the Zhongguancun Academy calls 'knowledge refinement.' AI companies worldwide, both Chinese and American, use this technique to improve efficiency and optimize models. The essence of this process lies in catching up development rather than creating original breakthroughs from scratch. Even Dean Boll from OpenAI asserts that distillation is not the main driver of China's recent achievements in AI.
If distillation is not the key issue, then what is? Experts from China and the US agree that the priority should be joint risk management, the development of common rules, and collective regulation, as the danger of AI safety recognizes no borders.