[The U.S. Proposes Allowing AI Giants to Legally Collaborate to Counter China's Model 'Distillation'] According to monitoring by Beating, bipartisan U.S. lawmakers have introduced the 'CATS Act.' The bill aims to provide limited antitrust protection for AI companies. Competitors would be allowed to share intelligence on model theft, distillation, and cyberattacks. In cases of national security risks, they could also jointly delay or restrict the deployment of high-risk models. 'Distillation' refers to extensively querying a powerful model and collecting its responses, then using that data to train a more cost-effective model. The bill's sponsors have explicitly identified China as the primary target. Previously, Anthropic, OpenAI, and Google accused Chinese labs of large-scale extraction of their model capabilities. Recently, the U.S. government also singled out Kimi K3 from Moonshadow. Before companies can jointly restrict models, they must provide written notification to the U.S. Department of Justice. They must also demonstrate that their actions are solely aimed at addressing security risks. Price collusion, market division, monopolization, and coordinated boycotts remain illegal. [Original Article Link]