Virginia Gov. Glenn Youngkin (R) vetoed a high-risk artificial intelligence regulation bill Monday, delivering a setback to state-level efforts to erect guardrails on the industry in the absence of federal action.
The bill represented the first major test of 2025 to enact sweeping AI regulation, following passage last year of a first-in-the-nation AI algorithmic discrimination law in Colorado.
“We are disappointed that the Governor chose to veto HB 2094, our bill on algorithmic fairness,” Del. Michelle Maldonado (D), the bill sponsor, said in a statement. “This legislation would have ensured that artificial intelligence could not be used to discriminate in Virginia based on housing, healthcare, education, marital status, and more.”
Youngkin’s veto is likely to energize AI regulation opponents who have been buoyed by the Trump administration’s warnings about “excessive regulation” of the technology.
Evidence of an AI regulation backlash in the states has emerged in recent months, with venture capital and free market groups blasting the idea of a patchwork of state laws and urging the Trump administration to preempt any that are passed.
Read more: Opposition to AI regulation is getting louder
The shift in the political landscape was underscored earlier this month when Texas Rep. Giovanni Capriglione (R) abandoned a bill that borrowed from Colorado’s framework. In its place, Capriglione introduced a new bill primarily focused on state government’s use of AI and ensuring that AI systems don’t engage in political viewpoint discrimination or incite criminality.
Capriglione’s change of heart and Youngkin’s veto suggest that a Republican version of the Colorado law is unlikely to take hold. Instead, red states and blue states may go more drastically separate ways when it comes to legislating AI.
It had appeared AI regulation might follow the path of consumer data privacy regulation in the states. Nineteen states have enacted data privacy laws, with most tracing their roots to an industry-approved framework that Virginia first passed in 2021. Consumer advocates have criticized that framework as too industry-friendly.
Similarly, consumer groups previously criticized the Virginia AI bill as riddled with “loopholes and exemptions” and appeared to welcome Youngkin’s veto.
“While we appreciated the intent of this bill, it had serious shortcomings that would have allowed many companies to evade the law in practice,” Grace Gedye, a policy analyst with Consumer Reports, said in a statement. “We want to make it crystal clear that legislators who want to protect their constituents — rather than big tech — should be setting their sights higher.”
Colorado’s 2024 AI law was widely viewed as the launch point for a state-by-state AI regulatory effort necessitated by congressional gridlock. That law focused on ensuring that AI systems used to make consequential decisions about people’s lives do not discriminate in areas such as employment, housing and health care.
The start of the 2025 legislative sessions spawned similar algorithmic discrimination bills in more than a dozen states.
The bills quickly drew detractors, including Adam Thierer of the libertarian R Street Institute, who warned of a “woke AI train … steaming full speed ahead in the States.”
Despite the growing friction, Maldonado was able to shepherd her bill — a narrower version of Colorado’s law — to passage along party lines in only 60 days. The final version sought to take into account feedback from both business groups and consumer advocates. Maldonado described it as a balancing act between “competing perspectives that are often diametrically opposed.”
Youngkin faced a steady drumbeat of calls to veto the bill, including from the U.S. Chamber of Commerce, which warned it would “have an outsized negative impact on small businesses.” The leader of another industry-backed group said the law would turn Virginia “from an AI hub into a ghost town.”
“The role of government in safeguarding AI practices should be one that enables and empowers innovators to create and grow, not one that stifles progress and places onerous burdens on our Commonwealth’s many business owners,” Youngkin said in a veto statement. “This bill would harm the creation of new jobs, the attraction of new business investment, and the availability of innovative technology in the Commonwealth of Virginia.”
Maldonado had signaled that if Youngkin signed the bill into law, she would be willing to continue discussions on how to improve it before it took effect in 2026. But she appeared to anticipate the veto last week in an interview with the public radio program 1A, saying she was prepared to reintroduce the bill next year.
“We will try again because we’ve already seen what happens when we don’t stay up with technology,” Maldonado said. “We just simply have to look … at what’s happened with social media platforms and the situation we’re in today.”
Algorithmic discrimination bills were also introduced this year in Arkansas, California, Connecticut, Georgia, Hawaii, Illinois, Iowa, Maryland, Massachusetts, Nebraska, Nevada, New Mexico, New York, Oklahoma, Rhode Island and Vermont, according to tracking by the government affairs firm MultiState.
The bills in Georgia, Hawaii, Maryland, New Mexico and Vermont have died or appear to be dead. The Connecticut bill faces the threat of a veto.
But proponents say they are not discouraged.
Matthew Scherer, senior policy counsel at the Center for Democracy & Technology, pointed to short legislative sessions and budget crises as barriers to passing bills this year. He also said there was never an expectation that red states such as Texas would pass “strong AI regulation” this year.
Scherer said bills he considers “strong” are still alive in California, Massachusetts and New York, and more could be introduced soon.
“I don’t think the sands are shifting, although that’s certainly the impression industry has tried to create,” Scherer said.
This story was updated to include a quote from the governor’s veto statement.