China slams US claims of 'industrial-scale' AI theft (www.aljazeera.com)

China has vehemently rejected recent allegations from the United States that its top AI companies, including DeepSeek and Alibaba, have engaged in "industrial-scale" theft of AI models from U.S. counterparts. The accusations were outlined in a joint advisory from three U.S. federal agencies, which claimed that these companies have been utilizing "billions of tokens" from U.S. models since at least late 2024, possibly with government oversight. The concept of token distillation, a common practice where less advanced AI models are trained on more sophisticated ones, was highlighted by U.S. officials as evidence of large-scale appropriation rather than normal industry practices. Significantly, China's response articulates its stance on technological sovereignty and self-reliance, asserting that the U.S. is trying to monopolize the AI industry through unfounded claims. The ongoing tensions underscore the competitive and strategic nature of AI development between the two nations, especially as they prepare for a high-level summit. With substantial investments in AI—nearly $300 billion—the stakes are high, and the allegations could have broad implications for international collaboration and policy in the rapidly evolving AI landscape.