U.S. Agencies Accuse Chinese AI Firms of Industrial-Scale Distillation of Claude, GPT, Gemini and Grok

Executive Summary

NSA, CISA, and FBI have issued an advisory accusing Chinese AI firms of conducting industrial-scale malicious distillation attacks targeting U.S. frontier AI models including Claude, GPT, Gemini, and Grok.

While model distillation differs from traditional malware hacking, the government advisory treats the operation as a systematic campaign for extracting restricted AI capabilities, with implications for AI espionage and future offensive capabilities.

Key Details

Industrial-Scale Distillation

The advisory describes what U.S. agencies characterize as industrial-scale malicious distillation attacks, where Chinese AI firms systematically query and extract capabilities from U.S. frontier models to train their own models.

Targeted Models

The advisory names Claude (Anthropic), GPT (OpenAI), Gemini (Google), and Grok (xAI) as targets of the distillation campaign.

Espionage Implications

While not traditional cyber espionage, the systematic extraction of frontier-model capabilities could ultimately strengthen Chinese military and cyber capabilities. U.S. officials allege the operation is designed to extract restricted AI capabilities for strategic advantage.

Government Response

The advisory represents a significant escalation in the U.S. government's framing of AI capability extraction as a national security threat, treating it alongside traditional cyber espionage.

Sources

• The Hacker News, 9 Sep 2026
• Reuters, 8 Sep 2026

Defensive research only. No exploit code or attack instructions.