The White House Is Trying to Figure Out What to Do About Chinese AI
The Trump administration is divided over how to address the rapid advancement of Chinese AI models, with the White House favoring stricter controls and the Commerce Department viewing such measures as impractical. Recent concerns focus on “distillation” attacks, where Chinese labs allegedly train their models using U.S. AI technology, prompting discussions of potential sanctions. No formal policy decision has been made yet, but the debate highlights growing tensions over AI security and export controls.
- ▪The White House and the Commerce Department are at odds about imposing restrictions on Chinese AI models that could rival top U.S. systems.
- ▪China’s Moonshot AI lab released the Kimi K3 model, which officials say was built by distilling Anthropic’s Claude Fable 5 model.
- ▪Anthropic accused Alibaba of conducting the largest known distillation attack, leading Treasury Secretary Scott Bessent to label such actions as potential IP theft and a sanctionable offense.
- ▪The Commerce Department, which oversees export controls, is considering incentives for U.S. labs to develop open‑weight models to counterbalance China’s capabilities.
Opening excerpt (first ~120 words) tap to expand
Hugo LowellPoliticsJul 22, 2026 5:00 PMThe White House Is Trying to Figure Out What to Do About Chinese AIThere’s a debate going on in the Trump administration over how to handle increasingly powerful Chinese AI models.Photo-illustration: WIRED Staff; Getty ImagesCommentLoaderSave StorySave this storyCommentLoaderSave StorySave this storyThe Trump administration is split over how to respond to the rapid rise of China’s leading AI models, according to multiple people familiar with deliberations.The debate is broadly divided between parts of the White House, which has pushed for stricter controls on Chinese AI that may soon rival the most powerful US models, and the Commerce Department, which has viewed those restrictions as unworkable.The conundrum has become the latest flash point for the…
Excerpt limited to ~120 words for fair-use compliance. The full article is at WIRED.