GoKawiil
Tech News
CISA, the NSA, and the FBI issued a joint advisory naming DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI as having harvested billions of tokens from Anthropic, OpenAI, Google, and xAI models since late 2024. The firms allegedly used fraudulent accounts, proxy networks, and automated failover systems to bypass rate limits and extract restricted reasoning data at industrial scale.
The NSA, CISA and FBI issued a joint advisory alleging that Chinese AI companies including DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z.AI extracted billions of tokens from US frontier models such as Claude, GPT, Gemini and Grok since 2024. The agencies say this data was used to train competing systems like DeepSeek's R1 and Moonshot's Kimi K2/K3, describing the practice as 'distillation activities at an industrial scale.'
U.S. security agencies have publicly accused six Chinese artificial intelligence companies of systematically using American AI models to train their own systems through a technique known as distillation. The practice allows a smaller model to learn from the outputs of a larger, more advanced one, effectively transferring capabilities without the original research investment.
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.
Anthropic's threat intelligence chief Jacob Klein says foreign actors, particularly in China, are using illegal methods to extract outputs from its Claude models and train cheaper rival systems, a practice known as distillation. He named Moonshot AI's Kimi K3 model as an example of technology he believes was built this way, describing an organized effort to evade Anthropic's account and access controls at scale.
A technical breakdown describes the 'efficient frontier' concept in LLM inference, distinguishing techniques that shift performance along a fixed tradeoff curve—like sacrificing latency for throughput or intelligence for speed—from techniques that expand the frontier itself, creating more overall efficiency. Examples cited include quantization, distillation, pruning, and reasoning-level adjustments, applied in the context of running large agentic coding models such as GLM-5.3 or Kimi K3 with KV cache reuse.