Utah Gov. Cox Calls on Congress to Set Frontier AI Safety Rules

Utah Gov. Spencer Cox wants Congress to create new federal duties for frontier AI, including reporting requirements, worker protections and outside evaluations. The request sharpens a Washington dispute over whether advanced AI needs new rules or whether the government already has the tools to control it.

Cox is not asking to halt AI development. He said the United States should lead in AI while keeping people in control, and said he is not calling for a moratorium or a shutdown. Instead, he wants a consistent national framework for the advanced systems generally described as frontier AI.

The rules Cox wants

His proposal pairs domestic oversight with a national-security restriction. Cox also urged the United States to block sales of necessary AI components to China, linking the safety framework to competition over the technology.

Cox’s proposed federal requirements

• Mandatory reporting of AI safety incidents.
• Protections for employees who raise AI safety concerns.
• Independent evaluations of cutting-edge AI models.
• Restrictions on sales of necessary AI components to China.

A cited warning, then competing responses

Cox cited recent unsanctioned cyberattacks against Hugging Face that he said were conducted by AI agents. He used that incident to argue that Congress should begin addressing AI risks. The account is part of Cox’s case for action, rather than a new federal finding.

The response from congressional leaders is not uniform. Senate Majority Leader John Thune said he was open to light-touch AI regulation. House Speaker Mike Johnson said he planned to meet with AI executives, while House Minority Leader Hakeem Jeffries called for immediate congressional action on AI risks.

An argument over who reviews powerful systems

Cox’s call for independent evaluations would make outside review a stated part of a federal frontier-AI framework. Whether Congress embraces that approach, a lighter-touch model, or no additional guardrails remains unresolved. For now, the governor’s proposal sets out a concrete alternative to leaving safety judgments primarily with AI companies.