On June 26, Democratic Representative Josh Gottheimer of New Jersey and Republican Representative John Moolenaar of Michigan introduced the bipartisan Cloud Security Act, a piece of legislation that would require AI companies to report suspected misuse of their platforms to U.S. federal authorities.
Supporters of the bill argue it would close a critical gap in existing U.S. export rules, which were crafted to stop geopolitical rivals such as China from acquiring cutting-edge advanced AI chips. (Fast Company has contacted both Gottheimer and Moolenaar to request additional comment on the proposal). While strict limits on the sale of these high-end chips are already enforced nationwide, the bill aims to give U.S. tech companies more flexibility to alert federal officials when they suspect foreign actors are leveraging their products to build sophisticated AI systems.
The cross-party bill is introduced as the Trump administration advances new plans to restrict public access to America’s leading AI models. In recent days, federal officials have worked out a deal with OpenAI that would dramatically limit which customers are allowed to access the company’s upcoming ChatGPT 5.6 model. This move comes only weeks after the U.S. Commerce Department pressured AI firm Anthropic to reverse the release of Fable 5, a version of its powerful Mythos model, citing national security concerns. The order effectively blocks Anthropic from selling its most advanced AI product commercially, and the company has complied by taking the model offline.
It is true that there are well-founded reasons to worry about the risks of making increasingly capable AI models widely available to the public, including the very real possibility that these tools will be exploited by bad actors for harmful ends. That said, the federal government’s growing push to restrict access to specific AI models also raises serious concerns. Even if these limitations offer some national security benefits, a broad, abrupt restriction on access also risks leaving many other organizations without the AI technology they need to protect their own operations and systems. In effect, these access rules also grant the federal government the power to set and pick the terms under which America’s top AI companies can serve the U.S. consumer market.
The Government Wants to Rein in Powerful AI, But There Are Downsides
