Key facts
- The U.S. Commerce Department initially ordered Anthropic to halt access to its advanced AI models for non-citizens, citing national security.
- China's Zhipu AI released the GLM-5.2 model under an open-source license shortly after the U.S. order.
- The U.S. has since withdrawn the export control order on Anthropic's models.
- China leads the U.S. in daily AI usage among white-collar workers, with 83% of Chinese professionals using AI compared to 65% of Americans.
- The U.S. leads in AI research and computing power, while China leads in AI adoption and open-source development.
The global race for artificial intelligence dominance has intensified, marked by a recent standoff between the United States and China. The U.S. Commerce Department initially ordered AI firm Anthropic to suspend access to its most advanced models, Fable 5 and Mythos 5, for non-citizens, citing national security concerns. This move effectively restricted access to these proprietary models worldwide.
In response, Chinese AI company Zhipu, operating as Z.ai, swiftly released its GLM-5.2 model under a permissive MIT open-source license. This action, along with the growing capabilities of other Chinese open-source models like DeepSeek's R1, highlights China's strategy of promoting widespread accessibility and customization of AI technology.
Following this development, Washington reversed its decision, withdrawing the export controls on Anthropic's models. Anthropic is expected to restore access to Fable 5 and Mythos 5. This strategic shift underscores the differing approaches to AI development: the U.S. emphasizes proprietary, controlled models for safety and monitoring, while China champions open-source solutions to accelerate adoption and establish global standards.
Data indicates a significant gap in AI implementation, with 83% of Chinese professionals actively using generative AI in their daily work, compared to 65% of Americans. Despite the U.S. leading in AI research and computing power, China's focus on open-source and practical application is rapidly closing the gap and potentially positioning it to dominate AI deployment.