Balancing safety and innovation

Editor's note: The following text is a transcript of a podcast story. To listen to the story, click on the arrow beneath the headline above.

MARY REICHARD, HOST: It’s Tuesday, the 3rd of June.

This is WORLD Radio and we’re so glad you’ve joined us today. Good morning, I’m Mary Reichard.

NICK EICHER, HOST: And I’m Nick Eicher.

First up on The World and Everything in It, moving the guardrails for artificial intelligence.

This week, the Senate takes up President Trump’s budget proposal. It’s more than a thousand pages on taxes and spending. But buried in the fine print is a clause that would prevent states over the next decade from enforcing their own AI laws.

REICHARD: Supporters say it’s about streamlining innovation. Others say it could affect efforts to protect children.

With writing and reporting from Harrison Watters in Washington, here is WORLD’s Anna Johnansen Brown.

ANNA JOHANSEN BROWN: Before tackling his newest project, Tim Estes developed tools to track down terrorists and traffickers. Now, he’s working on making AI safe for children.

ESTES: So many of the applications today that parents allow their kids to use, I don't think they fully realize, like, how much they have been engineered to addict them and to amuse their attention, and then how much they're left wide open into various harms.

Estes founded AngelQ to develop an AI-powered web browser. It adapts and filters the internet to children in age-appropriate ways.

But he knows not all AI companies have noble intentions.

ESTES: There comes a competition between companies to make it more and more engaging, and the kids are gonna get strip mined, you know, for their attention.

In May, Estes joined researchers from the Heat Initiative and Social Media Victims Law Center to publish a statement. It called for federal guidelines in developing AI tools.

ESTES: We have to put a line in the sand now, because they will keep going until someone stops them. Unfortunately, they've had a decade of driving forward, promising to self-regulate, and we've seen in social networking that is a joke.

Some of the statement’s proposals appear in a piece of legislation recently put forward in the Senate. In May, Tennessee Senator Marsha Blackburn reintroduced the Kids Online Safety Act. The bill aims to hold social media platforms and other internet companies accountable for harmful content, and push them to take steps to protect minors online.

But first, lawmakers have a different piece of legislation on the table, a budget reconciliation bill. It contains a decade-long moratorium on state laws that limit or restrict AI systems.

YOST: Ten years is an eternity.

Dave Yost is Attorney General for the state of Ohio. He joined nearly 40 other state Attorneys General calling on Congress to delete the AI policy.

They note that the bill would upend hundreds of pending or enacted laws about AI-generated content and data privacy.

Yost says Congress has a legitimate interest in regulating interstate commerce, but blocking states from regulating AI goes too far.

YOST: If Congress means to create a national standard, they need to do that, not just try to tell the states that they may not act to protect their citizens.

The White House sees it a little differently.

Venture capitalist David Sacks serves as President Trump’s top advisor on AI. Over the weekend, he talked about the risks of AI on his business podcast.

SACKS: I would say that China winning the AI race is a huge risk. I don't really want to see a CCP AI running the world. And if you hobble our own innovation, our own AI efforts, in the name of stomping out every possibility of X risk, then you probably end up losing the AI race to China because they're not going to abide by those same regulations.

In January, President Trump revoked a 2023 order from President Biden that he said put too much red tape around AI development. He’s since called for greater freedom for AI companies to develop models to rival those developed by China.

Tech developer Tim Estes says clearing red tape does not have to come at the expense of removing guardrails.

ESTES: I think America is big enough and thoughtful enough to actually have amazing, revolutionary AI that has to satisfy basic safety concerns. Just like having the best planes in the world with safety has not stopped us.

Even some generally in favor of deregulation are worried. Wes Hodges is Acting Director of the Center for Technology and the Human Person at the Heritage Foundation.

HODGES: Unless they're hiding something behind the curtain, it seems like it is taking away the protections without offering you know, that federal standard in return, and that does concern me.

Last month, Hodges joined Tim Estes and other groups calling for federal guidelines. This bill is not what he expected.

HODGES: Suddenly, anything that targets an algorithm or AI, depending on, you know, the definitions of this language that we're discussing, suddenly are would be stopped, would not be able to be enforced.

Senate Republicans are split on regulating AI. Texas Senator Ted Cruz has called for a light touch. Here he is during a Commerce subcommittee hearing in May.

CRUZ: Do we go down the path that embraces our history of entrepreneurial freedom and technological innovation? Or do we adopt the command and control policies of Europe?

Meanwhile, Missouri Senator Josh Hawley has said he will do everything he can to kill the AI provision before the bill goes to President Trump’s desk.

With Senators back from recess, committees will get to work reviewing the budget reconciliation bill this week. Ohio Attorney general Dave Yost expects Senators to receive a lot of feedback from their constituents.

YOST: AGS are talking to their Congress people, just like the lobbyists for big tech are talking with Congress folks.

Tim Estes hopes lawmakers will weigh the benefits of developing AI carefully alongside the costs of user safety.

ESTES: This is very strictly capitalism at the expense of kids and at the expense of adults too, running amok.

For WORLD, I’m Anna Johansen Brown.

WORLD Radio transcripts are created on a rush deadline. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of WORLD Radio programming is the audio record.