Congress pushes AI “misuse” reporting as access tightens

On June 26, a bipartisan group introduced the Cloud Security Act to require AI companies to flag suspected misuse to the federal government, aiming to close a loophole in export rules. But the push arrives as the Trump administration moves to restrict public a

A new proposal to rein in powerful AI models arrived with a warning built into the timing. On June 26, lawmakers moved to give the government faster visibility into suspected misuse—at the same moment the federal approach to limiting access to top American models is tightening.

Representatives Josh Gottheimer, a New Jersey Democrat, and John Moolenaar, a Michigan Republican, introduced the bipartisan Cloud Security Act. The bill would require AI companies to flag suspected “misuse” of their platforms to the government. Advocates say the legislation would close a critical loophole in current export rules designed to prevent rival countries like China from acquiring advanced AI chips.

Even with existing limits on chip sales, lawmakers want companies to be able to report foreign misuse more freely to federal authorities—so the risk doesn’t depend only on export controls that can miss what’s happening downstream.

The push is landing while the Trump administration advances plans to limit public access to leading American AI models. In recent days. federal officials worked with OpenAI on a deal meant to sharply restrict which customers are granted access to its upcoming ChatGPT 5.6 model. Weeks earlier. the Commerce Department pushed Anthropic to roll back its release of Fable 5. a version of its powerful Mythos model. citing security concerns. Anthropic observed the order and pulled the model offline, effectively preventing the company from selling its most advanced product.

Supporters of restrictions argue the stakes are real. They point to the danger that increasingly dextrous AI systems can be turned against national security. But the bill’s introduction also underlines a second fear—one that lands closer to how the U.S. protects itself when the government narrows access to the very tools that could be used defensively.

California Democratic Representative Sam Liccardo put it plainly when discussing the risk of taking defensive capability away from the institutions that would need it most. “Concern number one is if we’re going to take the defensive capability out of the hands of U.S. government, of local government, and of U.S. industry. and we’re going to tie our hands. at least one hand. behind our back. we better have a really good and clear justification for that because the downside risk is enormous. ” Liccardo said. “If these are really advanced models that are going to help us defend ourselves against the very advanced models to come. we’re nearly out of time.”.

Liccardo was one of four lawmakers who on June 18 sent a letter to the Commerce Department about the sudden new limits on Anthropic’s Fable 5. including questions about whether the agency bypassed normal regulatory processes to take the model offline and. in effect. targeted Anthropic. Responses were due by the close of business day on June 26, but Liccardo’s office says they never came.

He also argues the bigger issue isn’t only the ongoing attention focused on Anthropic. In his view, the concern stretches across the wider U.S. ecosystem—if adversaries build more powerful models while American institutions are denied access to leading U.S. systems, many may lack the technology they need to defend themselves.

There’s another pressure point lawmakers can’t ignore: oversight that is hard to measure. Liccardo warned that standards for how to enact these systems aren’t clear even to legislators. “If we have an export licensing regime without any review or substantial oversight until months or even years later. ” he said. “that oversight becomes worthless and meaningless in an industry where progress is measured by days and weeks.”.

Taken together, the Cloud Security Act and the administration’s access limits point in two directions at once. One approach tries to push more reporting responsibility onto AI companies. The other narrows the market for advanced models. For Liccardo and other lawmakers. the question isn’t whether national security matters—it’s whether the method is precise enough to protect the country without leaving other organizations unable to guard themselves.

And for now, the answers appear constrained by time. The bill would change how AI firms communicate suspected misuse to the federal government. while the administration’s steps continue to shape who can even reach the latest systems in the first place—an increasingly consequential choice as the debate moves from exports to access.

Cloud Security Act AI regulation national security export controls AI chips China OpenAI ChatGPT 5.6 Anthropic Fable 5 Mythos model Commerce Department Sam Liccardo Josh Gottheimer John Moolenaar oversight advanced AI models