President Donald Trump on Thursday signed an executive order urging states not to regulate artificial intelligence, although some in Congress are also considering temporarily blocking states from regulating AI.
Trump and some Republicans argue that the limited rules states have already adopted, as well as others that could follow, will stifle innovation and technology growth.
Critics of both political parties, as well as civil liberties and consumer rights groups, worry that deregulating the government would amount to a favor to big artificial intelligence companies, which enjoy little oversight and Trump's efforts extend beyond presidential power.
Here's what you need to know about state AI rules and what Trump signed.
What regulations exist at the state level and why?
According to the International Association of Privacy Professionals, four states—Colorado, California, Utah and Texas—have passed laws setting some rules for AI in the private sector.
These laws include limiting the collection of certain personal information and requiring more transparency from companies.
The laws are a response to AI, which is already permeating everyday life. Technology is helping Americans make important decisions, including who gets a job interview, apartment rent, mortgage, and even certain health care. But research has shown that he may make mistakes in these decisions, including giving priority to a certain gender or race.
“It’s not that AI makes mistakes and humans never do,” said Callie Schroeder, director of the AI and Human Rights program at the advocacy group EPIC.
“I can say to a person: “Hey, explain how you came to this conclusion, what factors did you take into account?” – she continued. “I can't ask the AI anything about it and I can't figure it out. And, frankly, half the time the AI programmers couldn't answer that question.”
More ambitious state AI regulatory proposals would require private companies to be transparent and assess possible risks of discrimination within their AI programs.
In addition to these more sweeping rules, many states regulate some aspects of AI: banning, for example, the use of deepfakes in elections and the creation of pornography without consent, or setting rules regarding the use of AI by the government itself.
What is the decree intended to do?
The executive order directs federal agencies to identify burdensome state AI rules and pressure states not to implement them, including by withholding federal funding, including for broadband, or challenging state laws in court.
It will also begin the process of developing a softer regulatory framework for the entire country that will take precedence over state AI laws.
It does not attempt to prejudge some of the laws passed by states, such as AI-related child protections and regulations on how state governments can acquire and use AI.
Trump argues that disjointed regulations across the 50 states are hampering the growth of AI companies and allowing China to catch up with the US in the AI race. The President also said government regulators are creating “Woke AI.”
Separately, House Republican leadership is discussing a proposal to temporarily block states from regulating AI, House Majority Leader Steve Scalise told Punchbowl News last month.
It remains unclear what the proposal would look like and what AI rules it would override.
TechNet, which advocates for tech companies including Google and Amazon, previously argued that the regulatory freeze would benefit smaller artificial intelligence companies still finding their feet and give lawmakers time to craft a nationwide regulatory framework that “balances innovation with accountability.”
Why Federal Regulatory Attempts Failed
Some Republicans in Congress have previously tried and failed to stop states from regulating AI.
Part of the problem is that the opposition comes from within their party.
Republican Florida Gov. Ron DeSantis said in a post on X last month that a federal law prohibiting government regulation of AI is “unacceptable.”
DeSantis argued that the move would be a “subsidy to Big Tech” and would prevent states from protecting themselves from a number of things, including “predatory apps targeting children” and “online censorship of political speech.”
A federal ban on states regulating AI is also unpopular, said Cody Wentzke, senior policy counsel in the ACLU's National Policy Advocacy Department.
“The American people do not want AI that is discriminatory, unsafe, or hallucinatory,” he said. “So I don't think anyone is interested in winning the AI race if it means the AI isn't trustworthy.”
What could happen next with the decree?
There is a good chance this will become part of the trial.
Democratic state attorneys general and governors have already challenged many of Trump's orders in court. No one has announced legal action on this issue yet, but it is an area of bipartisan interest.
In May, the attorneys general of 40 states and territories—Republicans and Democrats—signed a letter to congressional leaders calling on them not to pass a provision that would block government regulation of AI for 10 years.
Shatora Roberson, senior policy counsel for the Lawyers' Committee for Civil Rights Under Law, says it's clear in this case that the president doesn't have the power to override state laws.
“This is a matter of our democracy, and the president cannot, through executive order, simply override state laws without going through the democratic process,” she said.
California State Senator Scott Wiener, author of the landmark AI safety bill that was signed into law this year, called the order “blatantly corrupt and patently illegal.”
“If the Trump administration tries to enforce this ridiculous order, we will see them in court,” Wiener, a Democrat, said in a statement.
__
Associated Press writers Jeff Mulvihill and Tran Nguyen contributed to this article.
