U.S. agencies accuse China-based AI companies of distilling leading AI models

U.S. cybersecurity and intelligence agencies—the NSA, CISA, and FBI—have accused China-based artificial intelligence companies of systematically extracting proprietary functionalities and capabilities from U.S. frontier AI models through distillation attacks. The agencies described the activity as industrial-scale and said it involved using data from Claude, GPT, Gemini, and Grok to improve the companies' own models.