To beat back efforts to address artificial intelligence (AI) harms, earlier this summer opponents of meaningful regulation attempted to force through a decade-long moratorium on state-level AI regulation. While this effort failed, these opponents still parade around claims that leaving regulation to the states will result in a patchwork of AI regulation dominated by big blue states. For example, Adam Thierer argues that “California and New York will now take the lead in shaping national AI policy” and that “Their efforts to advance European-style technology mandates and the old Biden administration woke AI playbook will now have a much greater chance of moving forward in coming months and years.”
Ahead of Thursday’s House hearing titled “AI at a Crossroads: A Nationwide Strategy or Californication?,” we aim to cut through the noise and provide lawmakers and the public with the necessary facts. To do so, we analyze data from the Vanderbilt AI Law Lab (VAILL)’s database of artificial intelligence legislation. The database is “compiled from state legislative records, primarily utilizing the Legiscan API,” and is the best source of information on state AI legislation available. As of its most recent update on August 11th, the database contained information on 1,065 state-level bills active in 2025.
FACT: Very few artificial intelligence bills became law.
Just because bills are introduced does not mean that a “patchwork of laws that are impossible to comply with” becomes reality. Most bills – on any topic – have a slim chance of passing.
According to the VAILL database, only 219 artificial intelligence bills were signed into law in 2025. This represents only 21% of all artificial intelligence legislation that was active. The remaining 846 bills met the same fate as most legislation: they never made it to committee, stalled, failed a floor vote, or were vetoed.
FACT: Artificial intelligence legislation was active in blue states and red states.
It’s true that the states with the most active AI legislation in 2025 tended to be blue states. The top five states with the most bills were New York, Illinois, Texas (which isn’t a blue state), California, and Maryland.
However, looking only at the top states hides the fact that there was a significant amount of legislation in red states. Of the 1,065 bills, 448 (42%) were in states that Trump won in the 2024 election. Every state had artificial intelligence bills active in 2025.
FACT: More artificial intelligence legislation was signed into law in red states than in blue states.
As we mentioned above, very few bills actually become law. When we narrow our focus to these bills, a very different picture emerges.
Figure 1: AI Legislation Signed into Law by State, 2025
The above map shows the number of artificial intelligence bills signed into law in 2025. Many of the leading states are red states: Texas, Georgia, North Dakota, Montana, and Utah. While California and Illinois may rank 1st and 2nd respectively in terms of active bills, they rank 12th and 34th in terms of bills signed into law. Of the 219 bills signed into law, 136 (62%) are from states that Trump won in 2024.
FACT: Most artificial intelligence bills are not substantive and/or only apply to specific sectors.
In Trump Office of Science Tech Policy official Dean Ball’s own words, out of the 1,000+ bills only a few dozen are substantive “regulation”. Not every bill that contains the phrase “artificial intelligence” is regulation that would significantly affect AI companies. As former OpenAI safety researcher Steven Adler notes, some (if not many) of these bills simply create task forces or allocate funding.
The International Association of Privacy Professionals maintains a tracker of state AI governance legislation. Unlike the VAILL database (which includes everything), this tracker only includes bills that (1) only apply to private sector organizations, and (2) apply generally. In other words, this tracker lists the bills that would regulate AI in ways that the proponents of deregulation fear. As of its most recent update in July, the tracker only contained 56 bills (well short of 1,000+). And, of these 56 bills, only 6 have passed.
FACT: 1,000 bills is really not that much.
In the last legislative year, nearly 250,000 bills were introduced in state houses across the country. Assuming this number holds for 2025, this means the approximately 1,000 artificial intelligence bills only represent 0.4% of all state-level legislation this year. This is a drop in the bucket for a new technology that is allegedly supposed to upend our economy and society.