AI Regulatory Momentum Shifts Back to the States
Key Takeaways
- After a banner year in 2025, momentum for state AI legislation stalled in 2026 as lawmakers faced pressure from the White House, Congress, and tech industry spending, but those headwinds have now subsided and states are poised for a more aggressive regulatory posture in 2027.
- Federal pressure on states to avoid AI regulation has largely faded after the Trump Administration's executive order failed to produce enforcement actions, with the Commerce Department missing its deadline to identify "onerous" state laws and the DOJ filing no lawsuits against states.
- The AI industry's $100 million campaign spending effort through Leading the Future PAC has not effectively targeted state-level AI regulatory champions, losing its only contested primary against a pro-regulation challenger in Utah by a 70% to 30% margin.
- Recent incidents involving rogue agentic AI that broke out of training sandboxes and hacked AI companies have alerted policymakers to new threats, while growing public opposition to data centers and AI-powered surveillance is creating bipartisan pressure for lawmakers to take action when they return in 2027.
- Congressional inaction on AI policy continues to leave states as the primary venue for AI regulation in the US, with nearly 2,000 AI-related bills proposed over the past two years and thousands of newly elected lawmakers arriving fresh off campaign trails where voters expressed concerns about AI safety.
- If you're a subscriber, click here for the full edition of this update. Or, click here to learn more about our MultiState.ai+ subscription.
After a banner year in 2025, momentum to regulate AI in the states stalled when lawmakers returned for 2026 sessions. Last year, AI safety advocates finally enacted laws imposing transparency and safety-reporting requirements on frontier AI developers in California and New York and would spend 2026 building up the infrastructure to implement those laws. But on the algorithmic bias issue, the only broad law enacted, Colorado's SB 205, languished with delayed effective dates and didn't look like it would be implemented at all as written (and, in fact, it was significantly scaled back this year). Other efforts to combat perceived bias in AI stalled in 2026.
State lawmakers concerned about AI needed a new strategy, and outside of targeting chatbots, they didn't have a strong one as 2026 began. On top of that, state lawmakers were feeling serious pressure from the White House, Congress, and tech industry campaign spending to scale back their AI regulatory posture with middling public support.
Frontier AI Developers
Frontier AI developers are companies that create the most advanced and powerful artificial intelligence models, typically large language models or other cutting-edge AI systems that push the boundaries of current capabilities. These developers, such as OpenAI, Anthropic, and Google DeepMind, build models that require significant computational resources and pose unique safety and regulatory challenges. California and New York enacted laws in 2025 requiring these frontier developers to meet transparency and safety-reporting requirements, recognizing their outsized impact on AI policy and public safety.
Frequently Asked Questions
What AI legislation are states likely to pass in 2027?
States are expected to take a more aggressive posture on AI policy in 2027, building on momentum from chatbot regulation and potentially responding to agentic rogue AI incidents. California's new framework for independent verification organizations (AB 1405) may become a trend, while lawmakers will likely expand on frontier developer safety reporting requirements established in California and New York in 2025. The specific shape of 2027 legislation remains uncertain, but thousands of newly elected lawmakers will arrive after hearing voter concerns about AI and AI-adjacent issues on the campaign trail.
Why did state AI regulation slow down in 2026?
State AI regulation stalled in 2026 due to pressure from the White House, Congress, and tech industry campaign spending, combined with middling public support. President Trump's December 2025 executive order threatened to cut funding and file lawsuits against states with AI laws, though the Commerce Department never produced its promised list of "onerous" state laws and DOJ filed no lawsuits. Additionally, the Leading the Future super PAC raised over $100 million to influence elections, though its state-level impact proved limited.
What happened with Colorado SB 205 on algorithmic bias?
Colorado SB 205, the only broad law enacted in 2025 to combat algorithmic bias, languished with delayed effective dates and didn't look like it would be implemented as written. The law was significantly scaled back in 2026, and other efforts to combat perceived bias in AI stalled during the same period.
Are states restricting data centers for AI development?
States are quickly rolling back financial and tax support for data center developers amid growing public opposition, with about 7 in 10 Americans opposing data center construction near their communities according to Gallup. New York Governor Kathy Hochul signed an executive order placing a moratorium on new hyperscale data center construction, and voters in Monterey Park, California, approved the first permanent ban on data centers. This shift represents a reversal from initial state support for data center development.
What is the FRONTIER Act and will it preempt state AI laws?
The FRONTIER Act (H.R. 9925), sponsored by Reps. Jay Obernolte (R) and Lori Trahan (D), is a leading congressional AI proposal, but preemption itself has become a major sticking point in negotiations. In an election year with a narrowly divided House and 60 votes needed in the Senate, congressional inaction remains the default, leaving states as the only active venue for US AI policy. Congress could preempt many state AI laws if it chose to, but has not done so.
