US Agencies Accuse Six Chinese AI Firms of Copying Frontier Models
US intelligence and security agencies have accused six Chinese artificial intelligence firms of systematically copying advanced US AI models, potentially saving billions in development costs and accelerating their own progress.
The National Security Agency (NSA), Cybersecurity and Infrastructure Security Agency (CISA), and the Federal Bureau of Investigation (FBI) jointly released findings Tuesday alleging that six Chinese AI companies have been engaged in large-scale efforts to replicate capabilities from leading US AI models.
According to the agencies, companies including DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI have targeted prominent models such as variants of Claude, GPT, Gemini, and Grok. The alleged activity, which has been ongoing since at least late 2024, involves extracting functionalities that would typically require substantial investment in research and development.
The joint statement indicated that these actions allow Chinese AI firms to significantly shorten their development timelines and reduce financial expenditures associated with training frontier models. The agencies also noted that this activity likely occurred with awareness from the Chinese government.
This development raises concerns about intellectual property theft and the fair competition landscape in the rapidly evolving field of artificial intelligence. The coordinated statement from multiple US agencies underscores the perceived national security implications of these alleged practices.