GoKawiilUS agencies say Chinese AI firms systematically distilled ChatGPT, Claude, Gemini models
The NSA, CISA, and FBI issued a joint advisory alleging that Chinese AI companies, including DeepSeek and Moonshot AI, have run large-scale campaigns since 2024 to extract capabilities from leading US models like ChatGPT, Claude, Gemini, and Grok. The agencies say these firms used millions of queries routed through multiple accounts, APIs, proxies, and cloud providers to generate synthetic training data and build competing models such as DeepSeek's R1/R3 and Moonshot's Kimi-K2/K3.
GoKawiil's interpretation of the reporting above, not reported fact.
The advisory frames model distillation not as incidental copying but as a core strategic pillar of China's AI development, potentially undercutting the competitive advantage of US labs that spent billions building frontier models. It also highlights how evasive infrastructure—proxies, aggregators, and obfuscated metadata—makes such extraction hard to detect or block, raising questions about how US AI companies can protect their intellectual property going forward.
- US intelligence agencies formally accused Chinese firms of industrial-scale AI model distillation targeting ChatGPT, Claude, Gemini, and Grok.
- DeepSeek and Moonshot AI allegedly used millions of queries across multiple platforms to extract capabilities for their own models like R1/R3 and Kimi-K2/K3.
- Chinese companies reportedly used proxies, third-party aggregators, and obfuscated metadata to avoid detection while harvesting frontier model outputs.
Source: slashdot.org, 2026-09-08
Published there as: “Feds Accuse China of 'Systematic' Distillation of US AI Models”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.