Full Breakdown
U.S. Cybersecurity Advisory Accuses Chinese Artificial Intelligence (AI) Firms of “Aggressive, Malicious” Model Distillation; Beijing Denies Misconduct
9/9/2026, 10:16:44 PM
Core Event
U.S. federal agencies—the Federal Bureau of Investigation, the National Security Agency, and the Cybersecurity and Infrastructure Security Agency—issued a joint cybersecurity advisory alleging that six Chinese artificial-intelligence companies have conducted “industrial-scale, aggressive, malicious” distillation campaigns to extract restricted capabilities from frontier American models such as Anthropic’s Claude, OpenAI’s GPT, Google’s Gemini and SpaceXAI’s Grok. The advisory claims the activities have occurred “since at least late 2024” and were carried out “likely with Chinese government awareness.”
Background & Context
The accusations arrive amid heightened U.S.–China tensions over technology leadership and just before a scheduled summit between President Donald Trump and Chinese President Xi Jinping later this month. Earlier U.S. statements have framed model distillation—a technique that re-trains a smaller model on the outputs of a larger one—as a security risk when used to appropriate proprietary functionalities. Chinese officials have repeatedly countered that distillation is a standard, neutral practice employed worldwide, and that the U.S. allegations lack factual basis.
Data & Statistics
Industry tracker OpenRouter reported that Chinese large-language models recorded higher weekly usage than their U.S. counterparts for the 19th consecutive week, covering the period from August 31 to September 6. Four of the five most-used models globally were Chinese, with Tencent’s Hunyuan Hy4-preview topping the list at 14.7 trillion tokens, up 379 percent from the prior week.
Official Statements & Responses
- A spokesperson for the Ministry of Commerce described the U.S. attempts to suppress Chinese firms would meet “resolute countermeasures.”
- Scott Bessent, U.S. Treasury secretary, asserted that China “can never get ahead” of the United States in AI and linked the alleged distillation to broader concerns about Chinese military and cyber-attack capabilities.
Criticism & Opposition
Chinese industry insiders and scholars argue that the U.S. narrative conflates commercial competition with national-security threats, thereby creating a “double standard” that privileges American firms while vilifying Chinese developers. Liu Wei, director of the Human-Machine Interaction and Cognitive Engineering Laboratory at Beijing University of Posts and Telecommunications, contended that the U.S. is using security rhetoric to justify tighter restrictions on Chinese AI access.
Conversely, U.S. officials maintain that the alleged extraction of proprietary functionalities poses a direct risk to intellectual-property rights and national security, justifying defensive measures for American companies.
Conflicting Reports & Gaps
The U.S. advisory presents the distillation campaigns as “aggressive, malicious, and targeted,” yet Chinese sources describe the same technique as a legitimate, industry-wide method for model improvement. The advisory does not provide public evidence of wrongdoing, while Chinese ministries assert that the claims are “unfounded” and lack legal grounding. No Chinese company—DeepSeek, Alibaba, Moonshot AI, Z.ai, MiniMax, or StepFun—has offered a comment in response to the advisory.
What’s Next
The issue is slated to be discussed during the upcoming U.S.–China summit later this month, where both sides are expected to address AI governance, cooperation frameworks, and the broader implications of the advisory.
Verbatim Quotes
- “The Chinese distill our models and they can never get ahead of us,” — Scott Bessent, treasury secretary
- “Beating China — there is no day after tomorrow if China wins at this," Bessent said, read the report.” — Scott Bessent, treasury secretary
