FBI, NSA warn Chinese AI companies like DeepSeek and Alibaba are reportedly carrying out 'industrial-scale' distillation campaigns to boost their models
US AI companies should implement additional mitigations to curb these attempts, agencies warn.
The Cybersecurity and Infrastructure Security Agency (CISA), alongside the National Security Agency (NSA) and Federal Bureau of Investigation (FBI), have issued a joint warning to American AI companies about aggressive, malicious, and industrial-scale knowledge distillation operations conducted by Chinese firms like DeepSeek and Alibaba.
Knowledge distillation is a machine learning technique used to transfer learnings from a large "teacher model" to a smaller "student model," often for model compression and knowledge transfer. However, in this case, Chinese companies are allegedly using it to extract proprietary functionalities and capabilities from US AI models, including those based on Claude, GPT, Gemini, and Grok.
According to the agencies, these Chinese firms have been extracting billions of tokens from US frontier models since late 2024 by sending carefully designed questions and collecting the answers. The warning comes with recommendations for mitigation, such as detecting malicious prompts, deceptive responses, and unusual account behaviors.
The advisory also emphasizes the importance of cross-provider intelligence sharing among the US government, private industry, and allied nations to tackle the threat.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
