The fight over AI regulation in the US is moving from legislatures to courtrooms. After Congress failed twice to pass a law blocking state AI laws, President Donald Trump signed a sweeping executive order on December 11 that seeks to limit how far states can go in regulating the industry.
The order promises a “minimally burdensome” national AI policy and frames that approach as part of winning the global AI race. But states are under growing pressure to act on risks involving children, chatbots, data centers, bias, transparency, and frontier AI safety. That tension is now shaping one of the central technology policy fights of 2026.
The White House wants one lighter national approach
Trump’s executive order directs the Department of Justice to create a task force that can sue states whose AI laws conflict with his preference for light-touch regulation. It also directs the Department of Commerce to withhold federal broadband funding from states with AI laws the administration considers “onerous.”
That threat matters because some states may be unwilling or unable to risk losing federal broadband funding, especially where rural communities depend on it. Even if the administration loses in court, the uncertainty alone could discourage some state lawmakers from passing or enforcing AI rules.
James Grimmelmann, a law professor at Cornell Law School, expects the order to focus on a narrower set of state provisions. In his view, the likely targets are measures related to transparency and bias in AI, which he describes as issues that tend to be more liberal.
For tech companies and their allies, the order is a qualified win. They have argued that a patchwork of state rules would slow innovation, and some have built multimillion-dollar political operations to fight AI regulation.
States are still writing their own rules
Despite the executive order, several states are not stepping back. On December 19, New York’s governor, Kathy Hochul, signed the Responsible AI Safety and Education (RAISE) Act. The law requires AI companies to publish the protocols they use for safe AI model development and to report critical safety incidents.
On January 1, California introduced SB 53, the nation’s first frontier AI safety law. The law is aimed at preventing catastrophic harms such as biological weapons or cyberattacks, and the RAISE Act was modeled on it.
Both laws survived heavy industry lobbying only after being watered down from earlier versions. Even so, they represent a rare compromise between major technology companies and AI safety advocates.
If the White House challenges these laws, states such as California and New York are likely to defend them in court. Florida, where there are vocal Republican champions for AI regulation, could also become part of that resistance. Margot Kaminski, a law professor at the University of Colorado Law School, argues that the administration is on “thin ice” when trying to preempt legislation through executive action.
Congress is unlikely to settle the fight soon
Trump has said he wants to work with Congress on a federal AI policy. But the source article makes clear that Congress is not expected to deliver a bill this year.
The Senate killed a moratorium on state AI laws in July after it was inserted into a tax bill. In November, the House removed a similar effort from a defense bill. Those failures show how difficult it will be to create one national standard, even with pressure from the White House.
Brad Carson, a former Democratic congressman from Oklahoma who is building a network of super PACs backing pro-regulation candidates, says the executive order has made bipartisan agreement harder. He argues it has hardened Democratic opposition and exposed major divisions among Republicans.
Those divisions are visible inside Trump’s broader political coalition. AI accelerationists around the administration, including AI and crypto czar David Sacks, favor deregulation. Populist MAGA figures such as Steve Bannon warn about rogue superintelligence and mass unemployment. Republican state attorneys general also signed a bipartisan letter asking the FCC not to override state AI laws.
Child safety may create rare agreement
Public concern about AI is rising, especially around mental health, jobs, and the environment. In 2025, state legislators introduced more than 1,000 AI bills, and nearly 40 states enacted over 100 laws, according to the National Conference of State Legislatures.
One area where states may find broader agreement is child safety. On January 7, Google and Character Technologies, the startup behind Character.AI, settled several lawsuits with families of teenagers who killed themselves after interacting with the bot. A day later, the Kentucky attorney general sued Character Technologies, alleging that chatbots drove children to suicide and other forms of self-harm.
OpenAI and Meta are facing similar suits. Without AI laws already on the books, courts will have to decide how product liability laws and free speech doctrines apply to these new dangers. Grimmelmann says it remains an open question what courts will do.
States are expected to keep pursuing child safety laws because those measures are exempt from Trump’s proposed ban on state AI laws. On January 9, OpenAI made a deal with Common Sense Media to support the Parents & Kids Safe AI Act, a California ballot initiative. The measure would set rules for how chatbots interact with children by requiring age verification, parental controls, and independent child-safety audits.
Data centers and elections are next fronts
The fight is not only about chatbot behavior or model safety. States are also responding to backlash against data centers by considering bills that would require them to report power and water use and pay their own electricity bills.
Labor groups may push AI bans in specific professions if AI begins displacing jobs at scale. Some states concerned about catastrophic AI risks may pass safety bills modeled on SB 53 and the RAISE Act.
Money will intensify the fight. Leading the Future, a super PAC backed by OpenAI president Greg Brockman and Andreessen Horowitz, plans to support candidates who favor unfettered AI development. On the other side, super PACs funded by Public First, an organization run by Carson and former Republican congressman Chris Stewart of Utah, will support candidates who favor AI regulation.
The result is a messy contest over who gets to write the rules for AI: state capitals, Congress, courts, or the White House. In 2026, that question may matter as much as the technology itself.