The U.S. government is intensifying efforts to regulate powerful AI models deemed a national security risk. On June 26, Representatives Josh Gottheimer (D-N.J.) and John Moolenaar (R-Mich.) introduced the bipartisan Cloud Security Act, which would mandate that AI companies flag suspected “misuse” of their platforms to federal authorities.
The legislation aims to close a critical loophole in existing export controls designed to prevent rival nations, particularly China, from acquiring advanced AI chips. While current rules sharply limit chip sales, the bill would empower tech firms to more freely report suspicions that a foreign actor is using their products to develop advanced AI models.
This push coincides with the Trump administration’s broader efforts to restrict public access to leading American AI models. Federal officials recently worked with OpenAI on a deal that would sharply limit which customers can access its upcoming ChatGPT 5.6 model, signaling a tightening of both hardware and software oversight.
Critics warn such measures could stifle innovation and slow the U.S. AI industry’s global competitiveness. There are concerns that overreach may drive research abroad or create a black market for restricted technologies, undermining the very security goals these rules intend to achieve.