Senators on both sides of the aisle had asked the parliamentarian to find that the proposal to withhold funding for the broadband equity, access and deployment program from states that enact or enforce laws related to AI over the next 10 years violated the Senate's Byrd rule, which limits the inclusion of non-budgetary provisions in reconciliation bills.
While the parliamentarian found a similarly controversial injunction bond proposal and several other provisions of the budget bill to be subject to the Byrd Rule, meaning that these measures would need 60 votes to be included in the reconciliation act, the Senate adviser concluded that the proposed state AI moratorium wasn't covered by the rule, Sen. Jeff Merkley, an Oregon Democrat and ranking member of the Senate Budget Committee, said Sunday.
The decision, which was announced without offering further elaboration as to the parliamentarian's reasoning, means that the AI provision can stay in the budget reconciliation bill and can be approved with a simple majority vote.
The U.S. House of Representatives has already approved the budget reconciliation package containing the proposed AI moratorium, and U.S. Senate leaders have said they are hoping to send the bill to President Donald Trump's desk by July 4.
But the AI provision, which since arriving at the Senate has been tweaked to tie states' compliance with the proposal to their broadband funding, has been met with backlash by both policymakers and advocacy groups that have argued the federal government's inaction on overseeing AI systems makes movement by the states all the more important.
Democratic Sen. Maria Cantwell of Washington and Republican Sen. Marsha Blackburn of Tennessee, along with their respective states' attorneys general, were the latest to raise alarm about the provision during a press conference Thursday.
Cantwell, who confirmed that she had submitted materials to the parliamentarian supporting the exclusion of the measure from the budget bill, said she wasn't sure if her colleagues were fully aware of how broadly the proposed moratorium would sweep, noting that it has the potential to erase dozens of state laws that protect consumers from data misuse, bias and other harms that could arise from AI use.
Blackburn and the state attorneys general agreed that the proposal not only swept too broadly but would also cause harm to consumers if states weren't able to regulate rapidly emerging AI issues, especially if Congress continues to fall short in putting in place protections at the federal level.
"We need a national standard, but until we get it, it is our states that are filling in the gaps," Blackburn said. "You can't just say we're not going to regulate or enforce state laws for a period of 10 years, because think of what would happen to individuals in that period of time."
While neither senator commented on the parliamentarian's decision Monday, Consumer Federation of America, an advocacy group that was one of more than 140 organizations to oppose the moratorium in a letter to House lawmakers last month, spoke out against the provision being allowed to stay in the reconciliation bill and advance with a simple majority vote.
"This extreme measure is a clear gift to Big Tech at the expense of everyday people," Ben Winters, the group's director of AI and data privacy, said in a statement. "Barring states from regulating AI that harms their residents should be a complete nonstarter for lawmakers who want to put the people they represent over the profits of tech companies."
"This provision should have been stopped procedurally," Winters added. "It is a reckless overreach that is not only bad for consumers but for responsible innovators. This provision would leave the use of this technology unregulated and unaccountable."
Also on Monday, the Center for Democracy and Technology, which signed onto last month's letter as well, reiterated its opposition to the proposed 10-year ban on state AI regulation, focusing in particular on the problems with tying compliance to continued broadband infrastructure funding.
"Even if the AI moratorium were narrowed, it's still a bad policy that would force states to choose between accepting more funds to get communities online and protecting their residents from AI harms for 10 years," said Eric Null, co-director of the privacy and data project at CDT, adding that if the provision were enacted, AI companies in states that take more broadband infrastructure build-out funding would end up having "no accountability for their practices under either state or federal law."
--Editing by Emily Kokoll.
For a reprint of this article, please contact reprints@law360.com.
