US Accuses Chinese AI Firms of Mass Model Theft; Beijing Fires Back

The United States has accused leading Chinese artificial intelligence companies of systematically stealing American AI technology on an industrial scale, prompting a sharp rebuke from Beijing that the allegations are designed to preserve US market dominance.

Three US federal agencies—the FBI, National Security Agency, and Cybersecurity and Infrastructure Security Agency—issued a joint advisory Tuesday naming DeepSeek, Alibaba, and Moonshot AI as companies that have exploited American AI models to advance their own systems. The agencies claim these firms have distilled billions of tokens from US models since late 2024, likely with the Chinese government's knowledge.

The distillation dispute

At the heart of the controversy is a technique called distillation, where less sophisticated AI models learn from more advanced ones by processing their outputs. Tokens—the smallest units of text AI systems read or generate—serve as the training material in this process.

The US advisory characterized the scale and sophistication of Chinese distillation efforts as evidence that the practice forms "the critical core" of these companies' development strategy rather than a supplementary tool. American firms argue the volume crosses the line from standard industry practice into theft.

China's Commerce Ministry rejected the characterization entirely, telling the Associated Press that distillation is commonplace across the AI industry, including among US companies. The ministry accused Washington of seeking "a monopoly of the AI industry" and warned that "China will take resolute countermeasures" if the US uses distillation allegations to suppress Chinese firms.

Why it matters

The clash arrives weeks before Chinese President Xi Jinping's scheduled visit to Washington, where AI will feature prominently in discussions with President Donald Trump. The timing suggests both nations are positioning themselves ahead of high-stakes negotiations over technology competition. With China investing nearly $300 billion in AI and robotics as its next economic engine, and the Trump administration taking an increasingly hardline stance on Chinese tech advancement, the dispute signals deepening friction over who will lead the next generation of artificial intelligence development.

Diplomatic tensions rise

China's Foreign Affairs Ministry spokesperson Mao Ning called on the US to "refrain from making unfounded accusations or smears," asserting that "China's AI development is the result of high-level technological self-reliance and strength." She urged cooperation to promote AI development that is "open, inclusive, universally beneficial and oriented toward the common good."

US Treasury Secretary Scott Bessent struck a defiant tone Tuesday in Dallas, declaring that China can "never get ahead" of the US in AI. "The Chinese distill our models and they can never get ahead of us," Bessent said.

Xi's upcoming visit may include a large delegation of Chinese executives, according to Reuters, though the specific composition remains unclear. Previous summits have featured tech leaders—Xi brought Alibaba founder Jack Ma during his 2015 US visit, while Trump's May trip to Beijing included more than a dozen American executives, among them Tesla CEO Elon Musk.

The Chinese companies named in the advisory did not respond to requests for comment, according to Al Jazeera, which first reported these details.