The U.S. government accused six Chinese AI companies, including Alibaba and DeepSeek, of systematically extracting proprietary functions from American models through large-scale knowledge distillation, citing potential risks to national security ahead of a high-stakes meeting between Presidents Trump and Xi Jinping.
The U.S. government has formally accused six Chinese artificial intelligence companies, including Alibaba Group and DeepSeek, of systematically extracting proprietary functions and capabilities from American AI models through knowledge distillation, a process that could undermine U.S. technological leadership and national security. The allegations, detailed in a joint statement by the National Security Agency (NSA), Cybersecurity and Infrastructure Security Agency (CISA), and Federal Bureau of Investigation (FBI), come ahead of an expected White House meeting between U.S. President Donald Trump and Chinese President Xi Jinping.
Joint U.S. Agencies Detail Systematic Technology Extraction
The NSA, CISA, and FBI released a cybersecurity advisory on September 8, 2026, alleging that Chinese AI firms have engaged in knowledge distillation campaigns on an industrial scale since at least late 2024. The report specifically named Alibaba, DeepSeek, Moonshot AI, MiniMax, StepFun, and Z.AI, claiming they systematically extracted proprietary functions and capabilities from models developed by U.S. companies such as OpenAI, Anthropic, and Google. The advisory stated that Chinese AI companies were systematically extracting proprietary functions and capabilities from U.S. models.
Knowledge distillation, a widely used AI technique, involves training smaller models to mimic the outputs of larger, more complex systems. While the process is legitimate in academic and commercial contexts, the U.S. agencies alleged that Chinese firms have deployed it on a large scale to copy U.S. systems. The report cited millions of requests and interactions with U.S. models, including versions of GPT, Claude, Gemini, and Grok, to accelerate the development of their own AI products. The advisory noted that this approach significantly shortens the time required to develop advanced models and reduces training costs.
Alleged Government Involvement and Strategic Implications
The U.S. agencies suggested that the activities likely occurred with the knowledge of the Chinese government, though no direct evidence was provided. The accusation aligns with broader U.S. concerns about China’s growing AI capabilities and their potential military and cyber applications. The report warned that the extracted technologies could enhance China’s military and cyber warfare capabilities, posing a threat to U.S. interests and those of its allies.
The timing of the allegations has raised questions about their strategic intent. The U.S. is preparing for a September 24 meeting between Trump and Xi Jinping, during which AI safety and technological competition are expected to be key topics. The accusations may serve to pressure China ahead of the talks or highlight U.S. concerns about intellectual property theft. AFP reported that Washington has accused leading Chinese artificial intelligence companies of systematically extracting and copying the capabilities of American models amid an intensifying technological rivalry between the two countries.
Company Responses and Industry Reactions
Several of the accused companies have not publicly responded to the allegations. Moonshot AI declined to address the claims, while the other firms did not immediately respond to requests for clarification. The U.S. government has previously made similar accusations against Chinese AI developers, but this is the first time the allegations have been formalized in a joint agency statement.
The accusations have sparked debate within the tech industry. Some U.S. companies have expressed concerns about the competitive threat posed by Chinese AI models, which are often cheaper and faster than their American counterparts. In July 2026, dozens of startups urged the Trump administration to avoid imposing a complete ban on foreign AI models, arguing that restrictions should be limited to government use.
Broader Implications for U.S.-China Tech Relations
The allegations underscore the deepening rivalry between the U.S. and China in the AI sector, a critical domain for economic and military power. The U.S. has previously accused Chinese firms of using industrial espionage and unfair trade practices to gain technological advantages. In a reciprocal move, Beijing has accused American companies of using Chinese data to train their models, though these claims lack the same level of formal documentation as the U.S. allegations.
The situation highlights the challenges of regulating AI in a globally interconnected tech ecosystem. While knowledge distillation is a standard practice, the scale and intent of the alleged activities have drawn sharp criticism from U.S. officials. The BGNE report stated that distillation is a widely used method in which one AI model is trained to reproduce the output of a larger and more powerful model. However, U.S. officials claim that Chinese companies are deploying the technique on an industrial scale to copy U.S. systems unlawfully.
