Beijing Rejects US Allegations of Widespread AI Theft as Tensions Rise Ahead of Summit

The Chinese government has issued a forceful rejection of allegations made by the United States claiming that leading Chinese artificial intelligence enterprises have engaged in "industrial scale" intellectual property theft by exploiting American AI models.

The sharp dispute follows a joint security advisory issued by three prominent US federal agencies: the Federal Bureau of Investigation, the National Security Agency, and the Cybersecurity and Infrastructure Security Agency. In their published assessment, the agencies formally accused Chinese technology entities of systematically appropriating and utilizing American artificial intelligence systems to bolster and advance their own domestic capabilities.

According to the joint advisory, major Chinese companies—specifically identifying DeepSeek, Alibaba, and Moonshot AI—have harvested "billions of tokens" from advanced American models. The US intelligence and security bodies claimed that these activities have been actively taking place since at least late 2024, adding that the practices were likely being conducted with the direct awareness of the Chinese government.

In the realm of machine learning, tokens serve as the fundamental, smallest units of text that artificial intelligence models either read or generate during processing. The core of Washington’s complaint revolves around the technique known as model distillation, a process where smaller or less advanced AI models are trained on the outputs generated by larger, more sophisticated systems.

US officials argued in their joint warning that the immense volume and high level of sophistication observed in these operations point to distillation acting not merely as a supplementary tool, but as the critical core of Chinese model development. Because of the vast scale at which this data harvesting is reportedly occurring, American authorities and technology firms have characterized the practice as an unauthorized appropriation of intellectual property.

In response to the accusations, China’s Commerce Ministry released a statement firmly dismissing the claims and arguing that Washington is using the allegations as a pretext to maintain a monopoly over the global artificial intelligence sector. The ministry maintained that model distillation represents an established, commonplace technique utilized across the broader AI industry worldwide, including by US-based developers.

The Commerce Ministry warned that Beijing would not accept unilateral technological restrictions based on these accusations. Chinese officials stated clearly that if the US government suppresses Chinese artificial intelligence companies under the pretext of targeting distillation practices, China will take resolute countermeasures to defend its domestic industry.

China’s Ministry of Foreign Affairs similarly called on Washington to abandon its aggressive stance, demanding that US officials refrain from making unfounded accusations and smears against Chinese technological endeavors.

During a regular press briefing, Foreign Ministry spokesperson Mao Ning maintained that China's rapid progress in artificial intelligence is the direct result of high-level technological self-reliance, domestic innovation, and internal strength. Mao added that rather than leveling unilateral accusations, all international parties should instead strengthen cooperation to promote AI development that is open, inclusive, universally beneficial, and oriented toward the common good.

The sharp exchange of rhetoric comes at a delicate diplomatic juncture, arriving just weeks before Chinese President Xi Jinping is scheduled to travel to the United States. President Donald Trump is slated to host a comprehensive bilateral summit with Xi, where the two leaders are expected to hold formal discussions covering a wide array of topics, including developments in artificial intelligence.

Diplomatic observers note that the upcoming visit could see high-level industry involvement. President Xi may be accompanied to the US by a substantial delegation of Chinese corporate executives. While the precise roster of accompanying business leaders has not been formally released, past high-stakes meetings between the two nations suggest that executives from the artificial intelligence and broader technology sectors could be included in the traveling delegation.

Historical precedent highlights the role of executive delegations during bilateral visits between Washington and Beijing. President Xi previously included Alibaba founder Jack Ma in his official delegation during his 2015 trip to the United States. Similarly, when President Trump traveled to Beijing for a diplomatic visit in May, the US leader was accompanied by more than a dozen American corporate executives, including Tesla Chief Executive Officer Elon Musk.

The Trump administration has consistently pursued a hardline stance against the rapid development of Chinese artificial intelligence capabilities. While addressing an audience in Dallas, US Treasury Secretary Scott Bessent stated that China would remain unable to overtake American leadership in the sector, claiming that because Chinese firms distill US models, they can never get ahead of the United States.

Despite external pressure and US scrutiny, China continues to channel massive financial resources into its domestic technology sector. The Chinese government has poured nearly $300 billion into artificial intelligence, factory automation, and robotics, viewing these industries as the country's next primary economic engine.

The sprawling Chinese technology ecosystem now encompasses thousands of domestic enterprises developing humanoid robotics, embodied intelligence, advanced AI software, and autonomous systems. China remains locked in an intense race with the United States for global supremacy in artificial intelligence technologies, positioning its products to compete with lower manufacturing costs and strong commercial appeal. As the geopolitical friction persists, the Chinese companies specifically named in the US advisory—DeepSeek, Alibaba, and Moonshot AI—have not issued immediate public responses to the claims.