U.S. cybersecurity and intelligence agencies say six Chinese AI companies conducted industrial-scale distillation attacks on frontier AI models from Anthropic, OpenAI, Google, and xAI since late 2024. The advisory says the campaigns used fraud, proxies, and API abuse to extract model knowledge at scale, with DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI named in the report. #DeepSeek #MoonshotAI #Alibaba #MiniMax #StepFun #ZAI #Anthropic #OpenAI #Google #xAI #CISA #NSA #FBI
Keypoints
- Six Chinese AI firms are accused of large-scale distillation attacks on U.S. frontier AI models.
- The campaigns reportedly extracted billions of tokens through millions of requests.
- DeepSeek and Moonshot AI were identified as the most active offenders.
- The attackers used fraudulent accounts, proxies, and cloud services to evade limits and detection.
- CISA recommends stronger behavioral detection, response changes, and shared threat intelligence.