Another Ai Kerfuffle For Musk And Newsom?

Available free through April 11, POLITICO Pro’s Technology: California Decoded newsletter will explore how the Golden State is defining tech policy within its borders and beyond.

Quick Fix

— The lawmaker spearheading a major AI safety effort tells us why he’s feeling confident this year.

— Exclusive interview with California AG Rob Bonta reveals his thinking on Trump’s tech crusade.

— What a major tech lobby group is eyeing in Sacramento this year.

Hello and welcome to the first edition of California Decoded, the flagship daily newsletter from POLITICO’s brand-new California tech team. I’m Chase DiFeliciantonio, your anchor for today, bringing you the latest from the AI and automation beat, which will be my main focus. My partner Tyler Katzenberger will be bringing you all things tech policy, from social media to privacy debates. We’ll announce the final member of our team, who will cover Silicon Valley politics, tomorrow. Glad you're with us.

Send feedback, tips and story ideas to chasedf@politico.com and tkatzenberger@politico.com.

Driving the Day

ANALYSIS: AI SAFETY IS BACK — A marquee AI safety bill loathed by Big Tech, lauded by Elon Musk and slammed by leading congressional Democrats including Rep. Nancy Pelosi is back in Sacramento. The measure from state Sen. Scott Wiener is one to watch not just because of the controversy it generated last time around, but also for what its slimmed-down look says about who gets to regulate AI in the Golden State, and beyond.

Wiener, a San Francisco Democrat widely believed to have his eye on Pelosi’s seat when she retires, relaunched the bill Friday after the failure of the notorious, at least in tech circles, measure known as SB 1047 last year. But the battle lines have been redrawn with the retooled proposal, which for now is a far cry from the sweeping effort that generated headlines and a Big Tech backlash.

If Wiener wants to succeed this time, he will have to avoid a veto from Gov. Gavin Newsom, which felled his previous measure, and placate a governor who has shown little appetite for regulatory broadsides against one of California’s most-profitable industries. He will also have to convince and cajole Big Tech players like Meta to OpenAI to hold their fire, something he tried mightily — and ultimately failed to do — before.

Wiener appears to be taking a page out of the tech playbook of failing fast, resurrecting two key ideas from last year in the new bill to expand whistleblower protections for AI workers and to build out public computing resources for AI research.

“I would be very surprised if the bill drew meaningful opposition from tech,” Wiener told California Decoded on Friday after the bill language was first released. “But I’ve been surprised before.”

His revamped approach says plenty about who will be holding the reins in the effort to prevent runaway AI programs from using too much electricity or, as the naysayers fear, killing us all. Here’s a look at the key players Wiener faces with his latest effort on AI:

Musk: A long-time AI doomer, Musk somewhat grudgingly supported Wiener’s prior bill focused on pre-release testing of AI models in a surprise post last year. He did not put muscle or money behind his position, however. And with the new bill focused on making it easier for AI developers to sound the alarm from within their companies, it’s difficult to see Musk — who has shown himself to be no fan of whistleblowing and press leaks at Tesla — coming down on Wiener’s side again.

Newsom: The California governor not only vetoed Wiener’s SB 1047, but also charted an alternative path by appointing a who’s who of AI and legal experts to produce a report on how to best handle the safety risks posed by the technology. That writeup is expected imminently, within this first quarter of 2025. Wiener told California Decoded he hasn’t been privy to the group's work, but is open to expanding his bill to include the panel’s recommendations.

That puts him more in the waiting room than the driver’s seat with a smaller and more-focused measure potentially awaiting the input of Great Minds like Stanford’s Fei-Fei Li, whose oft-repeated moniker, The Godmother of AI, carries weight.

Democrats in Congress: Although Pelosi came out against Wiener’s prior proposal, whether she supports the slimmed-down version, or takes notice at all, remains to be seen. That intra-party scuffle was back when President Joe Biden was leading the White House instead of Musk — I mean, President Donald Trump. Under Biden, executive actions and Kumbaya international AI safety meetings were the order of the day. With Democrats firmly out of power in Washington, those in the party might be more inclined to notch wins on a key tech policy issue wherever else they can find them.

Big Tech: Last year, AI model makers OpenAI and Meta opposed Wiener’s more-sweeping measure, which drew some tepid support from Anthropic, the San Francisco maker of the Claude chatbot. Wiener is betting the new bill will draw less wrath from tech companies by focusing on whistleblower protections instead of the expansive vetting regime of the previous bill. It’s too early to know whether that happens, but we figured we’d ask anyway.

Anthropic declined to comment to California Decoded when asked about the bill, and OpenAI did not respond to questions. A representative for the San Francisco startup incubator Y Combinator, which opposed the previous version of the bill, similarly could not be reached for comment.

Among the Big Tech players last year, Meta strenuously opposed the bill, saying its testing rules created too much liability for startups using its free AI programs and hosting events to whip up opposition among Bay Area tech types. Meta spokesperson Jamie Radice told California Decoded that the company is reviewing the legislation. It’s not uncommon for industry to take their time reading through the legalese of a new bill before taking a side — or not.

Wiener said stripping out the testing provisions should allay much of the opposition. But replacing them with whistleblower protections is not a total olive branch in a tech industry where leaks can be tantamount to corporate treason that can cost significant cash and reputation loss.

None of the companies contacted by California Decoded gave a flat “no,” however, meaning tech still could come into the fold this time.

Happening Today

3:30 p.m. PT — The California Assembly’s Banking and Finance Committee will meet, including to hear testimony on Republican Assemblymember Phillip Chen’s bill aimed at saving cryptocurrency traders thousands of dollars on state-imposed licensing costs.

In the Courts

EXCLUSIVE: BONTA’S TECH PLAYBOOK — California Attorney General Rob Bonta isn’t losing sleep over Trump and Big Tech’s crusade against strict AI rules and data privacy protections, he told California Decoded in an exclusive, in-person interview today just hours before our launch.

It comes as Trump wages a pressure campaign against tech rules in Europe, whose approach often inspires California legislation.

“We expect that maybe Trump will want to sue. Maybe he won't. No idea,” he told us when asked whether the president might expand his pressure campaign to California. “We'll take him to court ... and as we usually do, we presume we have a high likelihood of success.”

The Trump administration has threatened tariffs in response to European taxes and fines on U.S. tech firms, with FCC boss Brendan Carr today lashing out at what he called Europe’s social media “censorship.”

Heard that line before? Us too. Tech industry groups — including Musk’s X — have filed a string of lawsuits arguing California’s social media and AI deepfake regulations violate the First Amendment.

Trump hasn’t implicated California, but both he and Musk have regularly made the state’s left-leaning policies a political punching bag.

Bonta told us his office isn’t waiting for Trump to pull the trigger.

“The laws that we engage on — either we sponsor or that we provide technical assistance on — we're doing deeper dives on all of those,” he said. “We know the playbook.”

Influence & Industry

GOLD RUSH — The liberal, pro-tech interest group Chamber of Progress is watching Golden State lawmakers like a hawk this year after opposing some of the state's most ambitious efforts to regulate social media and AI regulations last year.

CEO Adam Kovacevich shared a rundown today of the group’s top legislative fights in statehouses, and Sacramento features prominently.

Atop the watchlist are proposals to mandate warning labels on social media platforms, including Assemblymember Rebecca Bauer-Kahan’s AB 56, which would force platforms to display unskippable, 90-second health warnings to users at least once per day.

Bauer-Kahan also made the Chamber’s shortlist for her first-in-the-nation bid to stop tech companies from undercutting Hollywood creatives by using copyrighted works to train AI models without prior consent.

Kovacevich said the Chamber of Progress is monitoring scaled-back versions of legislation that failed last session, including Wiener’s revived AI safety effort and Assemblymember Cecilia Aguiar-Curry’s third try at sticking human safety operators behind the wheel of some autonomous delivery vehicles.

Byte Sized

— Apple is reportedly struggling to keep up in the AI race, despite a partnership with OpenAI (Bloomberg)

— Anthropic is now valued at $61.5 billion (CNBC)

— Alongside California, lawmakers in other states like Texas are grappling with energy demand for data centers (E&E News)

— John Bostic, who prosecuted Theranos founder Elizabeth Holmes, says her case is a reminder to Silicon Valley to be cautious and honest about new technologies (Mercury News)

With help from Nicole Norman

Have a tip, event or AI chatbot prompt to send us? Do reach out: Emma Anderson, California tech editor; Chase DiFeliciantonio, AI and automation reporter; and Tyler Katzenberger, Sacramento tech reporter.