Tuesday, October 14, 2025

Why California’s SB 53 would possibly present a significant examine on massive AI corporations

California’s state senate lately gave remaining approval to a brand new AI security invoice, SB 53, sending it to Governor Gavin Newsom to both signal or veto.

If this all sounds acquainted, that’s as a result of Newsom vetoed one other AI security invoice, additionally written by state senator Scott Wiener, final yr. However SB 53 is narrower than Wiener’s earlier SB 1047, with a deal with massive AI corporations making greater than $500 million in annual income.

I received the prospect to debate SB 53 with my colleagues Max Zeff and Kirsten Korosec on the newest episode of TechCrunch’s flagship podcast Fairness. Max believes that Wiener’s new invoice has a greater shot of turning into legislation, partly due to that massive firm focus, and since it’s been endorsed by AI firm Anthropic.

Learn a preview of our dialog about AI security and state-level laws beneath. (I’ve edited the transcript for size and readability, and to make us sound barely smarter.)

Max: Why must you care about AI security laws that’s passing a chamber in California? We’re coming into this period the place AI corporations have gotten essentially the most highly effective corporations on this planet, and that is going to be probably one of many few checks on their energy.

That is a lot narrower than SB 1047, which received numerous pushback final yr. However I feel SB 53 nonetheless places some significant laws on the AI labs. It makes them publish security reviews for his or her fashions. If they’ve an incident, it principally forces them to report that to the federal government. And it additionally, for workers at these labs, if they’ve considerations, provides them a channel to report that to the federal government and never face pushback from the businesses, though numerous them have signed NDAs.

To me, this seems like a probably significant examine on tech corporations’ energy, one thing we haven’t actually had for the final couple of a long time.

Techcrunch occasion

San Francisco
|
October 27-29, 2025

Kirsten: To your level about why it issues on the state stage, it’s necessary to consider the truth that it’s California. Each main AI firm is just about, if not primarily based right here, it has a serious footprint on this state. Not that different states don’t matter — I don’t need to be getting emails from the parents in Colorado or no matter —  but it surely does matter that it’s particularly California as a result of it’s actually a hub of AI exercise. 

My query for you, although, Max, is it simply looks as if there’s numerous exceptions and carve-outs. It’s narrower, however is it extra sophisticated than the earlier [bill]?

Max: In some methods, sure. I might say the principle carve-out of this invoice is that it actually tries to not apply to small startups. And principally, one of many important controversies across the final legislative effort from Senator Scott Weiner, who represents San Francisco, who authored this invoice, lots of people mentioned it may hurt the startup ecosystem, which lots of people take challenge with as a result of that’s such a booming a part of California’s financial system proper now.

This invoice particularly applies to AI builders which can be [generating] greater than $500 million [from] their AI fashions. This actually tries to focus on OpenAI, Google DeepMind, these massive corporations and never your run-of-the-mill startup.

Anthony: As I perceive it, if you happen to’re a smaller startup, you do must share some security data, however not practically as a lot.

It’s [also] price speaking in regards to the broader panorama round AI regulation and the truth that one of many massive modifications between final yr and this yr is now we’ve a brand new president. The federal administration has taken far more of a stance of no regulation and corporations ought to have the ability to do what they need, to the extent that they’ve truly included [language] in funding payments saying states can not have their very own AI regulation.

I don’t suppose any of that has handed up to now, however probably they might attempt to get that via sooner or later. So this could possibly be one other entrance wherein the Trump administration and blue states are combating.

Fairness is TechCrunch’s flagship podcast, produced by Theresa Loconsolo, and posts each Wednesday and Friday.

Subscribe to us on Apple Podcasts, Overcast, Spotify, and all of the casts. You can also comply with Fairness on X and Threads, at @EquityPod. 


Related Articles

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest Articles