The NSA, CISA, and FBI issued a joint advisory accusing Chinese AI companies, including DeepSeek and Alibaba, of conducting "industrial-scale" distillation campaigns targeting American frontier models2,3.
The U.S. cybersecurity and intelligence agencies described the activity as "systematic extraction" of proprietary functionalities and capabilities from models including Claude, GPT, Gemini, and Grok1. The advisory characterized distillation attacks as forming the "core" of these companies' AI development strategy.
The joint advisory, issued on Tuesday, represents a coordinated U.S. government response to what the agencies described as malicious copying of technology. Bloomberg reported that the U.S. government said Alibaba and DeepSeek have "systematically" siphoned AI models.
ANALYSIS The advisory names four distinct frontier model families as targets, spanning Anthropic, OpenAI, Google, and xAI, which frames the alleged distillation not as opportunistic but as a broad campaign across the leading U.S. model providers. A tri-agency advisory from NSA, CISA, and FBI carries a higher institutional weight than a single-agency statement, and the use of terms like "industrial-scale" and "core" of development strategy positions the accusation as a national security concern rather than a commercial IP dispute.