U.S. and China AI debate centers on distillation claims

Chinese AI distillation fuels U.S.-China rivalry as labs push faster, stronger models See why experts say China’s efficiency gains and chip limits are reshaping AI leadership

U.S. companies and officials have accused Chinese AI labs of relying on distillation, a method that uses outputs from more advanced models to train new systems. The practice has become a flashpoint in the broader competition between the two countries over AI leadership. Cohere CEO Aidan Gomez said Chinese models are advancing for reasons beyond distillation alone, arguing that some of the country’s latest systems are strong enough to outperform U.S. models on certain benchmarks. Other analysts also said China’s chip restrictions have pushed developers to build more efficient AI architectures. The article notes that Anthropic and U.S. agencies have described distillation as a major part of Chinese AI development, while China’s government has rejected those claims. It also points to the wider context of China’s tech sector, including faster innovation and increased domestic semiconductor development.