US Agencies Accuse Multiple Chinese AI Companies of Obtaining US Firms' Model Information Through "Distillation"

TechFlow reports that on September 9, Bloomberg stated that the U.S. National Security Agency, Federal Bureau of Investigation, and Cybersecurity and Infrastructure Security Agency alleged in a joint statement that Chinese artificial intelligence companies, including DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and ZhongAi Think Tank, have since at least 2024 been using the model "distillation" technique on an "industrial scale" to extract proprietary information from U.S. AI models and use it to train their own systems. The U.S. agencies simultaneously warned Silicon Valley developers to strengthen the protection of related outputs.