California, home of Silicon Valley, now has the most significant AI safety rules in the country.

From American Public Media, this is Marketplace Tech.

I'm Novosafo.

Welcome to Marketplace TechBtes. This week, in review.

This week, California passed a first-in-the-nation law setting sweeping new AI safety rules.

Also, YouTube settled a lawsuit brought by President Trump over account suspensions in the week of the January 6th Capitol riot.

And an AI-generated actress, quote-unquote, stirred up controversy in Hollywood and pretty much everywhere else.

For more on these stories, we spoke with Natasha Mascarinas, reporter at the information.

And we began with California's new AI safety law.

So what we're seeing that has been approved, it's called the Transparency and Frontier

Artificial Intelligence Act, and it requires AI companies to write, publish, and then

really comply with their own safety policies.

If they don't, they face penalties, but think things like, you know, a startup has to report

a safety incident to a state

official if their model gets stolen or misused. They also have protections for whistleblowers

for folks that are maybe working at these companies and want to share what is or isn't being

complied with. It also kind of lays the groundwork for a cluster called Cal Compute, which we can get into later.

But it's making the state actually get into the testing of AI models themselves versus just regulating them.

So interesting on many fronts.

And is that how it's ahead of the EU even, which passed a landmark AI safety rules, what, a year or two ago, I think?

Can you help us understand how much further the California rules are going? So it's the first time we're seeing

enforceable state level requirements for frontier AI labs. And so at least in within the

...