Safe and Secure Innovation for Frontier Artificial Intelligence Models Act

We can also expect any industry to fight tentative measures to impose real binding regulations.

This is exactly what AI companies did when SB 1047, a proposed California bill, was tossed on their desks.

Google opposed it directly, most concerning bill, well beyond federal measures, do not address industry concerns, unresolved problems, harm to R&D, and preferable approaches.

Meta opposed it directly too, significant concerns, we like safety too, deters innovation, disproportionate obligations, jeopardizes small businesses, not ready to move forward.

And then on the other side you had, you know, some corporate Democrats like Nancy Pelosi, uh, Zoe Lofgren.

Ro Khanna also opposed SB 1047, although he has since regretted that publicly.

The industry basically did not want to be regulated, but they especially did not want to have liability clarified because these AI companies operate in this kind of ambiguous zone where it's not clear who is liable when AI goes off the rails.

And with this Hugging Face hack, we have a situation where crimes were committed.

But to hack something, you need to have intent, and AI models can't have intent, it seems, under the law.

So it creates this, like, massive gap in [chuckles] the legal code.

And SB 1047 would not actually have covered that incident because the threshold was so high.

It was, like, half a billion dollars in damages and mass casualties, um, which, you know, dozens of people, I think, would have had to have died and, and nobody died from that situation, and the economic costs were, were much lower.

problems.

When I authored SB 1047, the first bill in the country to propose regulating AI, we were told we were promoting science fiction and that we were trying to regulate math.

How times have changed.

AI is unbelievably powerful.

Had a, a bill that was gonna hold the companies liable for mas- mass casualty events.

And California had the SB 1047, um, that Scott Wiener in- introduced and I think got, if I remember correctly, got really weakened over time.

Um, and like the companies pushed back on it.

They didn't want to have liability for mass casualty ev...

Look at California.

They recently passed a bill called SB 1047.

The core idea is to force the developers of the largest, most powerful AI models to implement incredibly strict safety protocols.

But the kicker is that the developers can actually be held legally liable if someone uses their model to cause catastrophic harm.

And then on the other side you had, you know, some corporate Democrats like Nancy Pelosi, uh, Zoe Lofgren.

Ro Khanna also opposed SB 1047, although he has since regretted that publicly.

The industry basically did not want to be regulated, but they especially did not want to have liability clarified because these AI companies operate in this kind of ambiguous zone where it's not clear who is liable when AI goes off the rails.

And with this Hugging Face hack, we have a situation where crimes were committed.

The FBI was reported, uh, to by Hugging Face.

But to hack something you need to have intent, and AI models can't have intent it seems under the law.

So it creates this, like, massive gap in [laughs] the legal code, and SB 1047 would not actually have covered that incident because the threshold was so high.

It was like half a billion dollars in damages and mass casualties, um, which, you know, dozens of people I think would have had to have died and, and nobody died from that situation, and the economic costs were, were much lower.

And so I just kind of kept working at it.

And I like to do a mix of like reporting on, you know, I've covered SB 1047, this AI safety bill in California, which was the first real flashpoint in the AI policy and regulation debates.

And I covered that full time for a few months and it ended up being like just a few, maybe a thousand words in the book.

But it gave me a really deep understanding of just how lobbying happens and how regulatory fights play out and the types of arguments that the industry makes and how it's just like every other regulatory fight.

Uh, it's interesting.

In, in some ways this feels a little like things have come full circle because, uh, you know, California passed this frontier AI law, SB53, that happened a year after this really heated discussion about another bill called SB1047.

Mm-hmm.

And one of the reasons that didn't pass is because there was a kill switch provision in there that people-

We had Brie Treece of, uh, Fathom to talk about the logic, um, and importance of IVOs on the podcast before.

Um, and this is just another indication of how this IVO concept is in-increasingly becoming sort of consensus at the state level and probably, um, of increasing interest at the, the federal level.

Um, in terms of, of what they put out, we know that after S-ten- SB1047, uh, the governor set up a similar sort of, uh, task force.

Right.

And I kept doing other reporting projects.

On top of that, one of the biggest runs of articles I did was covering SB 1047, which I think I was the only journalist covering that full time.

And that was in the summer of 2024.

And that was just really interesting to see the industry lobbying apparatus up close and doing a lot of quick turnaround and scoops and takes and That was, yeah, very informative to my kind of broader understanding of like what it will take to actually overcome the industry.