Article 6R4EP Gavin Newsom Vetoes Terrible AI Bill 1047, But Brace For Something Worse

Gavin Newsom Vetoes Terrible AI Bill 1047, But Brace For Something Worse

by
Mike Masnick
from Techdirt on (#6R4EP)
Story Image

I don't think I've ever seen quite as much hype about a state bill as California's SB 1047, a pretty terrible AI Safety" bill. Its supporters were a really weird combination of AI doomers, AI haters (not the same as the doomers), technically illiterate concern trolls... and a few people with legitimate interests about how to properly regulate technology.

The problem was that the few folks who had legitimate concerns and legitimate interests were more than willing to hitch their wagon to a bunch of numbskulls.

On the flipside, there were plenty of critics to 1047, and not all of them were great either. Some critics had legitimate concerns about how 1047 could stifle open source AI in particular, effectively locking in a few AI giants, but there were also some folks who went a little overboard on how terrible 1047 would be as well.

Either way, it was a bad bill. While California Governor Gavin Newsom has no problems signing absolutely terrible AI bills, he balked on this one. Of course, he's been signaling for months that this was what was going to happen. This is one of a few reasons we only wrote about 1047 once, even as it seemed to takeover the AI regulation media discussion. Since he always seemed likely to veto, why waste the energy?

On Sunday, he officially announced the veto and published a veto message, with an attempt to PR his way through the decision. Instead of just saying it was a bad bill, targeting the wrong things and written in a way that was not connected to the realities of how innovation works, he tried to still sound tough on AI. He said that effectively the bill was no good because it might make people think that other AI models not covered by the bill were safe:

By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology. Smaller, specialized models may emerge as equally or even more dangerous than the models targeted by SB 1047 - at the potential expense of curtailing the very innovation that fuels advancement in favor of the public good.

I... don't think that was the main problem of the bill, dude.

Elsewhere, his argument did make more sense, noting that any regulatory regime right now must be adaptable. This technology is still quite new, and regulating out of a place of fear of the unknown is a terrible way to regulate. But that's exactly what 1047 did. It assumed that fairy tales and moral panics were legitimate concerns about what AI tools could do... and also that there was some way to bureaucrat your way out of bad results.

Adaptability is critical as we race to regulate a technology still in its infancy. This will require a delicate balance. While well-intentioned, SB 1047 does not take into account whether an Al system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data. Instead, the bill applies stringent standards to even the most basic functions - so long as a large system deploys it. I do not believe this is the best approach to protecting the public from real threats posed by the technology.

My key takeaway from watching the debate of 1047 and other AI bills play out over the last few months is that a lot of people feel that (1) social media is bad, and (2) they missed a chance to regulate it when they should have, and (3) they don't want to do that with AI, and, therefore, (4) they need to overcorrect and aggressively regulate AI.

I think all four of those points are problematic and wrong. But we're living in the age of a raging tech moral panic, mostly because other shit is going horribly wrong on a societal level. But no one wants to do the hard work of actually fixing that shit, because that's hard and people are used to those problems. But tech is new and therefore, if we can just blame the tech and regulate the tech, surely we'll do something good.

It is not serious lawmaking. It's performative nonsense from unserious people.

But have no fear. As we've learned in the past, there will be a new version of 1047 next year (along with other terrible bills) and we'll get to have this kind of fight all over again.

External Content
Source RSS or Atom Feed
Feed Location https://www.techdirt.com/techdirt_rss.xml
Feed Title Techdirt
Feed Link https://www.techdirt.com/
Reply 0 comments