China has dismissed claims from three US federal agencies alleging that Chinese AI firms engaged in the industrial-scale theft of American AI models. Beijing has warned of 'resolute countermeasures' against any attempts to monopolize the AI industry.
- FBI, NSA, and CISA jointly accused Chinese firms of exploiting US AI models since late 2024.
- China's Commerce Ministry labeled the accusations as a move to establish a US monopoly over AI.
- The dispute centers on 'distillation,' which the US calls theft and China calls a common industry practice.
- Tensions escalate just weeks before President Xi Jinping's scheduled summit with Donald Trump.
China has issued a forceful rebuttal against claims by the United States that leading Chinese artificial intelligence companies have engaged in the 'industrial-scale' theft of American AI models. The diplomatic friction intensified after a joint advisory from the FBI, NSA, and CISA accused Chinese firms of systematically exploiting US-developed AI to enhance their own systems.
The US advisory specifically named companies such as DeepSeek, Alibaba, and Moonshot AI, alleging they have distilled 'billions of tokens' from American models. In the realm of AI, tokens are the fundamental units of text that models process and generate. US officials suggest these activities were carried out with the full awareness and likely support of the Chinese government.
Why This Matters
BozokMedia analysis shows that this confrontation transcends mere intellectual property disputes; it is a strategic battle for global technological hegemony. With China investing nearly $300 billion into AI and robotics to drive its future economic engine, the US is employing a hardline strategy to prevent China from closing the technological gap. The outcome of this 'AI Race' will likely dictate the global economic order for the next century.
China's Ministry of Commerce dismissed the allegations, arguing that the US is attempting to create a monopoly. The ministry pointed out that 'distillation'—the process of training smaller, more efficient models using the outputs of larger, advanced ones—is a standard practice worldwide, including within the US. However, Washington contends that the sheer volume and sophistication of China's campaigns indicate that distillation is the 'critical core' of their development, rather than a supplement.
"The weaponization of AI development narratives marks a new era of digital diplomacy where 'model distillation' is viewed through the lens of national security rather than scientific collaboration."
This escalation arrives at a critical diplomatic juncture, just weeks before President Xi Jinping is expected to visit the US for a summit with President Donald Trump. Reports suggest that Xi may be accompanied by a high-profile delegation of tech executives, mirroring past visits where business leaders played a pivotal role in bilateral discussions.
Adding to the tension, US Treasury Secretary Scott Bessent recently asserted in Dallas that China can 'never get ahead' of the US in AI, arguing that their reliance on distilling American models proves a fundamental lack of original innovation.
| Perspective | US Position | Chinese Position |
|---|---|---|
| Distillation | Industrial-scale theft of IP. | Standard industry practice. |
| Strategic Goal | Protecting national security/lead. | Technological self-reliance. |
| Government Role | State-sponsored exploitation. | Promotion of open and inclusive AI. |
Frequently Asked Questions
1. What is AI Distillation?
It is a machine learning technique where a smaller 'student' model is trained to mimic the behavior and performance of a larger, more complex 'teacher' model.
2. Which Chinese companies were named in the US advisory?
The US agencies specifically highlighted DeepSeek, Alibaba, and Moonshot AI.
