CISA, FBI back up US AI labs' China distillation accusations

CISA, FBI, and NSA issued a joint advisory accusing six Chinese AI labs—DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI—of industrial-scale distillation of US AI models, alleging China's government was likely aware. China's Foreign Ministry spokesperson Mao Ning rejected the claims, stating China's AI development stems from self-reliance.

Top US government agencies have accused six Chinese AI labs of illicitly extracting knowledge from US-made AI models to boost their own research, escalating allegations made by Anthropic and OpenAI. DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI were accused of large-scale distillation in a joint advisory https://www.cisa.gov/news-events/cybersecurity-advisories/aa26-251a?ref=thestack.technology by CISA, FBI, and NSA, which said China’s government was likely aware of the campaigns. It said “China-based AI companies are engaging in aggressive, malicious, and targeted distillation activities at an industrial scale” and claimed distillation was the “critical core” of their development process. While none of the six Chinese companies responded to a request for comment before publication, China’s Foreign Ministry Spokesperson Mao Ning told a press conference https://www.chinadaily.com.cn/a/202609/09/WS6aa113cfe4b06d4aa055d27d.html?ref=thestack.technology on Wednesday that the US should “ refrain from making unfounded accusations and smearing China.” She said: “China's development in artificial intelligence is the result of its high level of self-reliance and independent advancement in science and technology.”