Fast-moving innovation often leaves law and regulation in its dust, scrambling to catch up.
In November 2023, OpenAI CEO Sam Altman told Algorithmic Justice League founder Joy Buolamwini that there were a lot of laws restricting the misuse of artificial intelligence (AI), such as in medicine, but more are needed. “I think the trick in front of all of us — the technology developers, society, people — is going to be, now that we have this amazing new thing in the world, how do we get more of the good and less of the bad? And I think that is the story of technology and society,” Altman said.
Senate Bill 1047, which has already been passed by the Assembly Privacy and Consumer Protection Committee in Sacramento and is working its way through the state legislative chamber, would give the state attorney general power to “take legal action in the event the developer of an extremely powerful AI model causes severe harm to Californians or if the developer’s negligence poses an imminent threat to public safety,” according to a statement from the office of Senator Scott Wiener (D-San Francisco), who authored the bill with Senators Richard Roth, Susan Rubio, and Henry Stern, all Democrats.
Weiner says that, because the bill specifically targets only the largest models, it wouldn’t harm startups. “By taking a light-touch approach to regulation that focuses exclusively on the largest companies building the most capable models,” Wiener said, “SB 1047 puts sensible guardrails in place against risk, while leaving startups free to innovate without any new burdens.”
Despite Senator Wiener’s assurance that the bill specifically targets only the largest models, some in the AI startup community were nonetheless alarmed by SB 1047.
But some in the AI startup community were nonetheless alarmed by Senate Bill 1047. “It’s hard to understate just how blindsided startups, founders, and the investor community feel about this bill,” a16z General Partner Anjney Midha said on his firm’s podcast. “When it comes to policy-making, especially in technology at the frontier, our legislators should be sitting down and soliciting the opinions of their constituents — which in this case, includes startup founders. If this passes in California, it will set a precedent for other states and have rippling consequences inside and outside of the USA — essentially a huge butterfly effect in regards to the state of innovation.”
Wiener clapped back at the criticism, saying, “The hyperbole around SB 1047 notwithstanding, the bill is a focused approach designed to evaluate whether the largest new AI models create a risk of catastrophic harm. The bill only applies to huge models costing over $100 million to train. It doesn’t cover startups. It doesn’t ban open sourcing. It only authorizes the Attorney General to initiate enforcement proceedings and only in limited and extreme circumstances. It doesn’t require licensing or any form of permission from the government to train or release a model. The bill does require a safety evaluation for the largest models — evaluations that the large labs have already committed to performing — and common-sense safeguards to reduce the risk of catastrophic harm.”
But even large-scale AI players have voiced opposition to Senate Bill 1047. On June 25, Meta released a letter to Weiner saying that though it shares the goal of safe and responsible AI development, the bill places liability on the wrong participants, on the original model developers rather than on developers of harmful AI tools elsewhere in the AI ecosystem. “The bill’s fundamental flaw is that it fails to take this full ecosystem into account and assign liability accordingly, placing disproportionate obligations on model developers for parts of the ecosystem over which they have no control. This is not to suggest that regulation should exempt developers altogether. The problems lie in the types of obligations imposed on developers, which overwhelmingly center on certifying specific outcomes (an unrealistic task) rather than requiring governance and accountability measures,” wrote Meta’s vice president and deputy chief privacy officer for policy, Rob Sherman.
In May, American Enterprise Institute Senior Fellow Will Rinehart listed 31 bills before the California state legislature that would regulate AI, so regardless of the fate of Senate Bill 1047, AI will keep moving forward and legislation will continue scrambling to catch up.
