After years of the U.S. lagging behind other countries in AI regulation, President Trump signed an executive order last week aimed at limiting individual states from creating their own AI regulations.
According to the White House announcement and fact sheet, the executive order will assess and challenge state laws that "stifle innovation," with the intention of eliminating a "patchwork" of state regulations and creating one national standard. The details of that national standard, however, are still lacking.
"There is no universal American policy for AI," said Jack Gold, founder and principal analyst at J.Gold Associates. "What Trump is doing is saying, 'OK, we don't know what that policy is going to be, but you states can't get in the way.'"
Each state has unique needs and wants around AI regulation, Gold added. As states have waited for federal regulation, they've proposed and enacted their own laws to restrict the discriminatory use of AI in employment or address consumer protections related to AI systems, for example.
But these individual state AI laws can result in hodgepodge approaches that are difficult to manage. While state regulations can provide more nuance, that approach isn't sustainable at scale, said Michael Leone, practice director and principal analyst at Omdia, a division of Informa TechTarget.
"I think most could agree that, in the long term, it's simply unsustainable to think that we should be [regulating AI] at a state level," Leone said. "The market's too quick."
I think most could agree that, in the long term, it's simply unsustainable to think that we should be [regulating AI] at a state level.
Michael Leone, practice director & principal analyst, Omdia
The U.S. is a global leader in AI development, he added. It has been aided by states that foster companies building innovative AI software without much oversight or regulation. But state-level AI regulations require those companies to monitor each applicable law.
"If you were to look at this at scale, let's imagine there are 50 states with 50 different rules or laws," Leone said. "If I'm a Microsoft, Google or Amazon, I'm opening myself up to so many different legal battles that it's going to be detrimental to progress and adoption."
President Trump's executive order takes the stance that states can hinder AI innovation with laws that create unnecessary red tape and impede U.S. advancement in its AI battle with China.
"A carefully crafted national framework can ensure that the United States wins the AI race, as we must," the executive order said. "Until such a national standard exists, however, it is imperative that my Administration takes action to check the most onerous and excessive laws emerging from the States that threaten to stymie innovation."
The executive order calls for the creation of an AI Litigation Task Force, which is responsible for challenging state laws that are "inconsistent with the policy set forth." The Secretary of Commerce will also publish an evaluation of existing "onerous laws" at the state level, the order said.
For both Leone and Gold, however, this edict from President Trump doesn't do enough.
"The problem I have with this kind of [executive order] is all he's saying is, 'Don't do anything. I'm not going to do anything, but you can't do anything either,'" Gold said.
This creates a period of uncertainty, Leone said, as AI developers, states and enterprises wait for the federal policy.
"There are a lot of impacts to not having regulations from a political standpoint," Leone said. "It sets a dangerous precedent to prioritize innovation at all costs in order to 'maintain competitiveness with China.'"
Who wins and loses?
Who stands to benefit from the new executive order, and who loses?
"The winners are clear," Gold said. "It's the AI companies, because they don't want regulation."
While this can result in faster AI development, it also poses risks. To date, most of the major AI players have made minimal efforts to self-regulate and make their products safer, taking only small steps that help them avoid lawsuits and negative press, Gold said. As a result, these companies push for innovation without always considering the ramifications, such as child safety, bias and copyright infringement, among other issues.
"The losers are going to be, ultimately, the end users like us who are either using [AI] directly or are impacted by companies using AI inappropriately or dangerously," Gold said.
The losers are going to be, ultimately, the end users like us who are either using [AI] directly or are impacted by companies using AI inappropriately or dangerously.
Jack Gold, founder & principal analyst, J.Gold Associates
Leone also noted the friction between AI efficiency and the democratic safeguards around that innovation.
"If I'm a developer of services that are founded in AI, this is good in terms of innovation," Leone said. "It's not great in terms of safeguards, and the safeguard piece is really important."
Meanwhile, both Leone and Gold said they expect state regulators to resist the order at some level, implementing new AI restrictions or maintaining existing ones and letting the courts decide. That could take years to resolve.
Time will determine when the U.S. finally establishes a universal AI framework -- and what that policy includes. But Leone said he hopes the order stimulates some regulatory progress.
"It'd be nice if [the order] starts opening the eyes of Congress a little more to take action and push some cross-party agreement on what regulation should look like and why it should look like that," Leone said.
Jennifer English is editorial director for TechTarget's AI & Emerging Tech group.
