The Trump administration is once again targeting Chinese AI: new restrictions on the horizon
Washington has reignited discussions on how to block Chinese open-weight artificial intelligence models. According to my information, this round of debates is directly linked to the recent release of Kimi K3 by Moonshot AI — a model that appears to have caused serious concern in the corridors of power.
No official decisions have been made yet, but the range of tools under consideration is quite broad. These include placing Chinese AI developers on export restriction blacklists, which would effectively block any interaction with them by American companies without special permission from the Department of Commerce. Also under discussion are issuing public warnings about potential security threats from such models and introducing restrictions in the area of government procurement.
Regulatory Club or Competitive Struggle?
Interestingly, the administration previously refrained from issuing strict decrees that would place direct responsibility on American companies for using Chinese AI, fearing it would slow down its own industry. Now the rhetoric has shifted toward more subtle but no less effective tools.
Former White House senior AI advisor Dean Ball, now working at OpenAI, expressed the opinion that the best strategy for the Trump administration would be to create a "major regulatory risk" for using Chinese open-source models. He emphasized that Kimi K3 is indeed a high-quality model, and its success cannot be dismissed as mere data distillation. However, the president's AI advisor David Sacks harshly criticized this approach, calling the use of regulatory uncertainty as a competitive tool "absolutely unacceptable" and noting that it could only benefit the duopoly of OpenAI and Anthropic.
Double-Edged Sword
Meanwhile, Beijing is not sitting idle either. Chinese authorities are actively discussing with Alibaba, ByteDance, and Zhipu the introduction of their own restrictions on the export of advanced AI technologies and training data. Alibaba has already banned its employees from using Anthropic products, citing security risks. This reminds me of a classic escalation in high-tech fields, where each side tries to protect its assets, but ultimately the entire open innovation ecosystem suffers.
At this point, U.S. agencies have not presented specific evidence of hidden vulnerabilities in Kimi K3, and Moonshot AI itself has not yet been added to any lists. However, the very fact that these debates have resumed suggests that we are on the brink of a new phase of technological confrontation, where regulatory risks could become as significant as technical barriers.
Expert opinion: The AI market is entering a phase of "cold war." If the U.S. imposes restrictions on the use of open-source models, it could radically change the development landscape, forcing startups and researchers worldwide to choose between technology accessibility and legal safety. In the long term, this will inevitably slow down the pace of global progress in AI.