Congress’ yearslong struggle to force Big Tech to protect children online is now tangled in the fight over regulating artificial intelligence, adding new obstacles to an elusive legislative deal.
What was once a battle to take on Google, Apple, Meta, X and TikTok has now expanded to companies such as OpenAI and Anthropic.
Lawmakers eager to crack down on social media companies, particularly regarding children’s safety, sound more cautious about overregulating AI while the industry is still developing.
“We must address catastrophic risk without ceding ground to China or restricting Americans’ free expression,” said Sen. Ted Cruz, Texas Republican and chairman of the Senate Commerce Committee, which has jurisdiction over both issues.
Mr. Cruz made the comment after President Trump issued an executive order in June focused primarily on artificial intelligence, cybersecurity and national security, according to the White House. A separate executive order issued in December 2025 addressed AI innovation and state preemption.
The executive order sought to strike a delicate balance between mitigating the harms of AI and not stifling U.S. advancements in the industry.
Lawmakers are considering similar concerns as they look to place guardrails on AI as part of a legislative package to protect children online.
The preemption debate stems from a desire to give AI companies regulatory certainty as they advance their innovations.
“There’s been a lot of discussions back and forth trying to resolve the issue of preemption, how you respect states’ rights and still create a uniform standard, so that you don’t have 50 different laws you’re dealing with,” Senate Majority Leader John Thune, South Dakota Republican, told The Washington Times.
“That’s the objective of the White House, and a lot of our members,” Mr. Thune said. “They continue to work on how to thread that.”
Mr. Thune said the Kids Online Safety Act, or KOSA, a bipartisan bill that the Senate passed on a 91-3 vote last Congress as part of a combined package that also included the Children’s Online Privacy Protection Act update, could be “an engine that pulls the AI bill along.”
Sen. Marsha Blackburn, Tennessee Republican and co-author of KOSA, has been spearheading negotiations with the White House to do just that.
The goal, a Blackburn spokesperson said, is “to finalize legislative text of an AI preemption package that includes protections for kids, creators, and communities through the Senate version of KOSA, the NO FAKES Act, and age verification requirements.”
The NO FAKES Act is a bill to protect against unauthorized digital copies of a person’s voice or likeness amid a rise in AI copycats. The measure, led in the Senate by Sen. Chris Coons, Delaware Democrat, and co-sponsored by Ms. Blackburn, advanced out of the Senate Judiciary Committee in June.
Sen. Richard Blumenthal, a Connecticut Democrat and co-author of KOSA with Ms. Blackburn, opposes adding AI preemption to the child-focused legislation.
“Preemption should not be a part of it, period,” he said.
The House went through a similar partisan struggle as it worked on its own child online safety legislation.
The Energy and Commerce Committee marked up a package in March called the Kids Internet and Digital Safety Act, or KIDS Act, that included a version of KOSA and seven other bills, including safeguards for minors interacting with AI chatbots.
Democrats opposed the package in committee, partly because it included language preempting states from enacting or enforcing similar laws.
Energy and Commerce Committee Chairman Brett Guthrie, Kentucky Republican, and ranking member Frank Pallone, New Jersey Democrat, struck a deal on a new version of the KIDS Act that would ensure all states adhere to the federal standards in the bill at a minimum but allow them to regulate beyond that.
The House passed the bipartisan version June 29 on a 267-117 vote.
Mr. Pallone told The Times that if Ms. Blackburn plans to modify her bill to preempt state regulation of AI, that would be a “nonstarter.”
“In our bill, we were trying to make sure that it was a floor, not a ceiling,” he said. “Now she wants to preempt AI? That’s terrible.”
Mr. Guthrie, however, welcomed the Senate to go further than the House in that regard.
“The problem is if you have 50 different standards, it’s difficult for companies to operate,” he said at a recent Punchbowl News event.
Although Ms. Blackburn is leading negotiations with the White House, it is ultimately up to Mr. Cruz to decide what legislation to advance through his committee.
Mr. Cruz has promised a committee markup this summer on KOSA and a few other children’s online safety bills. He has said he wants to include AI bills in the mix, but he has been coy about specifics.
“Legislating in this area is complicated and fraught with peril,” Mr. Cruz said during the Senate Judiciary Committee markup of the NO FAKES Act.
He supported advancing that bill out of committee but said he wanted additional free speech protections to be added before a floor vote.
Mr. Cruz told Politico he was vetting bills for the Commerce Committee markup based on what can get bipartisan support.
“This markup is designed to move legislation that has a real chance of passing into law,” he said.
Sen. Maria Cantwell of Washington, the panel’s top Democrat, said, “Nobody on the committee knows what Sen. Cruz is proposing.”
She wants to advance the Senate version of KOSA but said, “Certainly we don’t need to tie it to bad AI policy.”
Despite disagreements over AI preemption, senators are mostly united in believing that their version of KOSA is better than the House’s because it includes a “duty-of-care” provision.
The duty of care would hold social media companies liable if they fail to adhere to the bill’s standards for platform designs that aim to prevent harmful online content from reaching children.
House Republicans have widely panned the duty-of-care provision as inviting a court challenge for regulatory overreach that could limit free speech.
“We can’t pass duty of care,” Rep. Morgan Griffith, Virginia Republican and member of the Energy and Commerce Committee, told The Times. “It is seen as being the entryway for a European-style controlling [of] the speech on the internet.”
Disagreements over the duty of care and AI preemption make it difficult to find a compromise on children’s online safety legislation that will please both parties and both chambers, Mr. Griffith said.
“But we’ve got to do something, because right now, we’ve got nothing.”
Contact the author
Please read our comment policy before commenting.