California's AI Safety Legislation: Governor's Veto and Future Implications

The governor's veto disappoints many, highlighting urgent AI safety concerns that demand attention and action.

### Disappointment Over Governor's Veto of AI Safety Legislation
* 😞 The governor's veto of the proposed AI safety legislation has been met with disappointment, as it was a focused effort to **address catastrophic risks from powerful AI models**.
* ⚖ The legislation aimed to require major tech companies to **conduct safety evaluations on their models before release** and to have the authority to **shut them down if necessary**.
* 🔍 Despite the industry's commitment to safety evaluations, the desire to **codify these requirements into law** was deemed **reasonable and necessary**.
* 🤔 The governor's rationale for the veto, citing concerns about a **false sense of security** and **exemptions for smaller models**, has raised questions about the **clarity of his position**.
* 👥 The establishment of a working group to consult with experts on future guardrails has been announced, though there are **concerns about the suitability of some members**.
* 📊 Despite intense lobbying from large tech firms, the legislation received strong support in both the Senate and Assembly, highlighting the **importance of addressing real risks associated with AI**.
* 🌍 The debate sparked by this bill has led to a **broader national and international conversation about AI regulation**, underscoring California's role in **leading safety initiatives in technology**.
* 🚀 While the bill's failure is disappointing, it has opened up **essential discussions** that will continue as stakeholders work towards **effective solutions for AI safety**.

Your reaction? Disappointment. But why? Less safe? Sure. Yes. The governor's veto is very disappointing. This is a very thoughtful, focused piece of legislation that would have addressed catastrophic risk from only the largest, most powerful models. And what it did was very basic. It would require the largest AI companies and largest tech companies to perform safety evaluations on massive, powerful models before they released them, and to be also to be able to shut down those models if something goes awry. They've all committed to perform these safety evaluations. But instead of just trust us, we want to put it in statute. So there was nothing extreme or unreasonable about this. And the veto, unfortunately, sends a very bad message about the future of safety. And, Senator, what do you make of Governor Newsom's rationale for vetoing this bill? He said that he's worried it would provide a false sense of security, that

smaller models could be exempt. We also saw Nancy Pelosi put out a statement about how she's happy the bill has been reached. What do you make of their rationales? Well, respectfully, to the governor, I'm focused on at least the governor acknowledged that these risks are real because some of the opposition throughout were saying, oh, this is just science fiction. It's not science fiction. These are very, very real, tangible risks that these models could help facilitate. So the governor did acknowledge that, which is good. I didn't fully understand the governor's veto message. It said while the bill is too broad, but the bill is also too narrow. So the veto message, it's unclear what the governor actually wants and will we'll see with this working group that he's setting up what comes out of that. But I and many others are committed to really moving forward to support innovation because I can make the world a better place.

But trying to get ahead of the risks posed by this very powerful technology. Right. And talking about that working group. So the governor has said that he's consulting with outside experts, including Faith Lee and some others, to potentially come up with, you know, guardrails, maybe even new legislation, it sounds like, in a future session. What do you make of that? Do you think those are the right people for this group? Some people have brought up that Feeley does have funding for a startup from Andreessen Horowitz, the VC firm, which has been very vocal against SB 1047. What do you make of that group? Are they the right people to be addressing this? Well, they also he put out on there, who's a former California Supreme Court justice, who I admire greatly and I have a lot of respect for for Lee. She's a incredibly smart, talented person. And it's also important that folks who

are technologists who supported the bill also be involved, because we had some of the greatest minds in AI who were supporting the bill. I hope that they are included as well. Right. Senator, what of the influence, though, of those that were against what of the lobbying? Did it make an impact? Well, it did not have significant impact in the legislative process. The legislature, with very strong votes in both the Senate and the Assembly, passed. So Democrats overwhelmingly supported this legislation, despite the intense lobbying by large tech companies and by large venture capital firms. You know, the governor went through his own process and I am confident was hearing a lot, both support and opposition and ultimately made his decision. And I strongly disagree with the governor's decision. But again, at least he acknowledged that these risks are real and that they're not made up. Will you re-introduce SB 1047 in the

next session? We're going to evaluate next steps. We certainly want to engage with the governor and with others. This issue is not going away. It's important. Congress has shown no signs of acting. Congress could not even get it together to pass an election deepfakes bill before this election. So California, as the heartland of tech and AI innovation, has a responsibility to lead on safety. So, Senator, while this bill didn't pass, it did spark a national, even international conversation about A.I. legislation. What do you think the impact has been? Has it surprised you how big this debate over this bill became and how much it divided the tech community? Yeah, I knew this bill would be controversial. Any tech regulation bill is. I did not anticipate the scale of the conversation. But I have to say it's always disappointing when a bill fails or gets vetoed. But sometimes you have to introduce a

big controversial bill to spark a long overdue conversation. I've seen that happen in some of the housing work that we've done. Where you pass it, you introduce a huge bill. It doesn't pass, but it moves the conversation forward. And I think the same is true here. The fact that the governor issued a very, very lengthy and detailed veto message. The fact that he acknowledged that the risks were real and the fact that he committed to forming not just forming a working group, but saying specifically who's going to be on it and what it's going to focus on. That's a big step forward. And so, yes, it's disappointing that this veto happened, but this issue isn't going away and we're going to get the job done.

Newsom Blocks Contentious AI Safety Bill in California

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.

Bloomberg Technology

Disappointment Over Governor's Veto of AI Safety Bill

The governor's veto of the AI safety bill is disappointing, as it aimed to require large tech companies to perform safety evaluations on powerful models and have the authority to shut them down if needed.

The veto sends a negative message about the future of safety, and the governor's rationale for the veto is seen as unclear and contradictory.

Despite the setback, there is a commitment to continue supporting innovation and addressing the real risks posed by these AI models.

The governor is consulting with outside experts, including Faith Lee, to establish guardrails and potentially new legislation for AI technology.

There is debate about the group's composition, particularly Lee's connection to a VC firm opposing SB 1047.

Despite intense lobbying from tech companies, the legislation passed with strong support from Democrats, and the governor's decision has been met with disagreement.

safety regulation is urgent, with Congress showing no action on the issue.

California, as a tech and AI hub, has a responsibility to lead in this area.

The controversial bill's introduction sparked a significant national and international conversation about A.

legislation, surprising the senator with its scale and the division it caused within the tech community.

A big controversial bill can spark important conversations, even if it doesn't pass.

The governor's detailed veto message and commitment to forming a working group are seen as positive steps forward.

Despite the disappointment of the veto, the issue remains a priority and efforts to address it will continue.