Senate Republicans have modified a controversial provision in their sweeping tax bill that would restrict states from regulating artificial intelligence, after facing sharp criticism from lawmakers on both sides of the aisle — including within their own ranks.
The original House-passed version of the legislation included a blanket ten-year ban on any current or future state laws aimed at regulating artificial intelligence technologies. But after backlash from state legislators, digital rights advocates, and some federal lawmakers, Senate Republicans restructured the language in a bid to preserve the provision while making it more palatable under Senate rules and political pressure.
Under the new language unveiled Thursday night, instead of an outright ban, the bill now proposes denying federal broadband funding to states that attempt to regulate AI over the next decade. The change reframes the issue as a budgetary condition, attempting to pass constitutional and procedural muster while maintaining a deterrent effect on state legislatures.
“These provisions fulfill the mandate given to President Trump and Congressional Republicans by the voters: to unleash America’s full economic potential and keep her safe from enemies,” said Sen. Ted Cruz, chairman of the Senate Commerce Committee, in a statement defending the AI-related amendment.
The move is believed to have been buoyed by members of the Republican Party who said the federal government had no business overriding the rights of individual states to govern emerging technologies in their own jurisdictions.
Rep. Marjorie Taylor Greene, R-Ga., who had voted in favor of the bill in the House, publicly walked back her support after learning of the provision.
“We should be reducing federal power and preserving state power. Not the other way around,” she wrote on social media. Greene admitted she had not read the AI provision before casting her vote.
Several GOP state lawmakers also denounced the measure, warning it sets a dangerous precedent for centralized overreach. Some stated that the federal government should not be using funding as a weapon to coerce states into surrendering their legislative authority — especially in matters as consequential and fast-moving as AI.
The backlash reflects a deepening rift in Republican ideology: while many in the party support limited government and states’ rights, others are prioritizing a unified federal approach to technological innovation, particularly in sectors where the United States faces stiff competition from global rivals like China.
Adding to the controversy, digital safety experts and civil liberties advocates warned that stripping states of their ability to legislate AI oversight leaves millions of Americans vulnerable. They point to risks associated with bias in AI systems, facial recognition surveillance, data privacy, and automated decision-making tools increasingly deployed by businesses and law enforcement.
Despite these concerns, several leading AI executives have lobbied for a centralized regulatory framework.
However, no bipartisan agreement has emerged in Congress over how to regulate AI at the federal level. Democrats have proposed stronger consumer protections and guardrails on algorithmic use, while Republicans have focused more on fostering innovation and limiting what they view as government overreach.
To help ensure the tax bill survives the fast-track budget reconciliation process, which allows it to pass with only a simple majority, Senate Republicans altered the AI provision to tie it to federal spending. They argue the measure qualifies as budget-related rather than a purely policy-driven initiative, by threatening to withhold broadband infrastructure funds from non-compliant states. Sen. Cruz said he plans to make this case to the Senate parliamentarian, who is expected to issue a determination in the coming days.
In addition to the AI language, the tax package includes a renewal of key tax cuts from Donald Trump’s 2017 legislation, new business incentives, expanded commercial spectrum access for telecom firms, and sweeping cuts to social safety net programs.
Senators are aiming to pass the legislation by the end of June, though the future of the AI provision remains uncertain. However, the revised approach has only intensified the debate over how much authority states should retain in shaping the rules for one of the most disruptive technologies of the 21st century.
