This is a linkpost for https://www.wsj.com/tech/ai/californias-gavin-newsom-vetoes-controversial-ai-safety-bill-d526f621
The Democrat decided to reject the measure because it applies only to the biggest and most expensive AI models and doesn’t take into account whether they are deployed in high-risk situations, he said in his veto message.
Smaller models sometimes handle critical decision-making involving sensitive data, such as electrical grids and medical records, while bigger models at times handle low-risk activities such as customer service.
For those that have been following this: Is he serious, or is this just lip service and he's blocking it because he was lobbied by people in the tech industry?
(Not super well-informed) My guess is that it's 95%+ lip service. He doesn't seem like someone with scruples or object-level opinions about most things other than what gives him power.
One new thing to me in that thread was that the California Legislature apparently never overrides the governor's vetoes. I wonder why this is the case there and not elsewhere.
Newsom's press release and veto message include much more detail and suggest "it's too weak" is not the actual reason.
Reasons mentioned:
Suggests Newsom is going to be very hostile to any legislation that is designed to deal with X-risk concerns, and that he, frankly thinks they are bullshit. (I personally am also pretty skeptical of X-risk from AI, but I don't want nothing done given how bad the risk would be if it did manifest.)
Good analysis of this from PauseAI:
I don't want to presume to paraphrase their analysis into one phrase, but if I were forced to, it would seem to be that there was a lot of pressure on Governor Newsom from powerful AI companies and interests, who also threatened to ruin the bill's sponsor Scott Wiener.
Still a pity that he couldn't resist the pressure.
It's kind of pathetic, but this is the reality of politics today. With their money, they really can either make or break a politician, and we voters are not smart enough to avoid being taken in by their negative advertising and dirt-digging.
It's clear that we need a much stronger movement on this. The other reason he was able to veto this bill is that the vast majority of people do not agree that AI poses a major / existential risk, and so they do not insist on the urgent action we need.
This seems pretty bad news from an AI safety perspective :/
Any chance to override his veto, or get a similar bill passed soon?
Some people have talked about trying to get something like that as a ballot initiative, since initial polling suggests that something like this might be very popular with the general public.
Yep, numbers ranged from 60% to 80% support for approving SB 1047, and it was impressively bipartisan, too.
To do a [citizen-initiated] ballot initiative, you stand on the street and ask passerby to sign your petition. Perhaps it would be possible to simultaneously build a mailing list of interested passerby to attend protests and such. That could translate the poll numbers into a stronger street presence.
I talked to people who seem to know about this and ~0 modern ballot initiatives are done this way. You need >500,000 signatures, which is a massive logistical undertaking that's not going to be enacted by a couple of volunteers.
More context:
TechCrunch article (not paywalled like the WSJ article)
Response from Senator Wiener on Twitter + discussion
Thanks for sharing!