China has formally rejected allegations from the United States that its leading artificial intelligence laboratories are engaged in the systematic theft of American AI capabilities, dismissing the claims as “groundless” amid intensifying technological rivalry between the two superpowers.
In a statement issued on Wednesday, China’s Ministry of Commerce characterized the U.S. position as “a typical case of using the crackdown on distillation as a pretext to achieve industrial monopoly.” The response follows a cybersecurity advisory released by a U.S. federal agency on Tuesday, which accused prominent Chinese AI firms—including DeepSeek and Moonshot AI—of conducting what it described as “industrial-scale knowledge distillation campaigns” to replicate the functionality of American models.
Knowledge distillation is a widely recognized technique in the AI industry whereby a smaller model is trained to emulate the outputs of a larger, more sophisticated system. While the method is commonly used across the global sector, Washington contends that Chinese enterprises have deployed it on a massive scale to illicitly copy proprietary U.S. technology.
The U.S. advisory specifically named DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI, alleging that these companies engaged in millions of exchanges and requests to extract capabilities from American AI systems developed by OpenAI, Anthropic, Google, and xAI. According to the advisory, firms adopting this strategy experience “significantly shorter AI development timelines and reduced financial expenditures in training a frontier model.” The document further asserted that such activities have been ongoing since at least late 2024 and were “likely conducted with Chinese government awareness.”
Agence France-Presse (AFP) reached out to the named companies for comment. Moonshot AI declined to respond, while StepFun could not be immediately contacted. The remaining firms did not issue public statements at the time of reporting.
The diplomatic friction emerges ahead of an anticipated meeting later this month between U.S. President Donald Trump and Chinese President Xi Jinping at the White House, where artificial intelligence is expected to feature prominently on the agenda.
Washington’s heightened scrutiny of Chinese AI development comes as American models from providers such as Anthropic and OpenAI—while generally regarded as more powerful—face growing competition from Chinese alternatives that offer comparable performance at a lower cost. This competitive dynamic has placed pressure on the Trump administration to formulate a response that balances national security concerns with the realities of a globalized technology market.
In July, a coalition of dozens of startups submitted a letter to the Trump administration urging caution against any outright ban or stringent restrictions on foreign-made AI models. The letter argued that such measures should only be contemplated in the context of government use of the technology, rather than as a blanket prohibition affecting the private sector.
Beijing, for its part, has turned the tables on Washington by accusing the United States of using Chinese data and examples to train its own AI models. At a regular press briefing on Wednesday, Chinese foreign ministry spokeswoman Mao Ning reiterated this stance, urging the United States to “refrain from making unfounded accusations against and smearing China.” She emphasized that the progress of artificial intelligence in China is the result of independent innovation and open collaboration, and that China remains committed to the responsible development of AI technologies in accordance with international norms.
The exchange underscores the growing strategic importance of artificial intelligence in U.S.-China relations, with both nations vying for leadership in a field that is expected to shape economic competitiveness, national security, and global influence for decades to come. As the two governments prepare for high-level talks, the dispute over AI distillation methods is likely to remain a contentious point, reflecting broader tensions over technology transfer, intellectual property, and the rules governing the digital economy.