US cybersecurity agencies have accused six Chinese artificial intelligence companies of conducting large-scale distillation attacks against leading American AI models. According to the statement, DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI extracted billions of tokens through millions of requests made to frontier systems operated by Anthropic, OpenAI, Google, and xAI.
To execute these attacks, the companies reportedly distributed API requests across fraudulent or shared accounts while using proxies to bypass geographic restrictions, usage limits, and detection mechanisms. This systematic approach allowed them to harvest vast amounts of training data from protected systems without authorization. None of the accused companies have responded to the allegations.