California Gov. Gavin Newsom on Sunday blocked a bill aimed at preventing large-scale AI programs from causing disaster, saying the law would have created a “false sense of security.”
His decision came after weeks of negotiations and competitive lobbying efforts by major tech firms, celebrities, billionaires, and AI developers.
The legislation, Senate Bill 1047, introduced by state Senator Scott Weiner (D-San Francisco) back in May, would have required companies spending more than $100 million in computing resources to create a basic AI model, or $10 million in resource spending. fine-tune the basic model, conduct safety tests, hire independent auditors to review the model annually and take “reasonable care” to ensure that the model does not cause mass casualties, more than $500 million in damage to physical or cyber infrastructure, or outsourcing in the observation of a person to commit the most serious crimes.
It also ordered the developers to build the killing machine into suitable models that would allow them to be shut down quickly and gave the government attorney the power to sue the developer for infringement and, in extreme cases of injury, seek damages of up to 10. a percentage of the cost of training the model.
Newsom in recent weeks has signed a series of bills into law that address the immediate, ongoing harm caused by AI systems, including bills that criminalize the creation of non-consensual pornographic images and require the AI models that generate them to flag their content for easy recognition. . But SB 1047, which would have applied only to the richest and most influential AI companies, has become a hotbed of debate over AI legislation in recent months.
Companies including Meta, Google, OpenAI and Anthropic strongly oppose the bill, saying it will undermine innovation and hurt small businesses, despite the rules only applying to the hundreds of millions of companies that must be used to train AI systems, for example, Meta , the bill would unfairly penalize developers of basic models for disasters caused by downstream users and discourage the creation of open source models because developers fear being held accountable for how others use their products.
In his veto statement, Newsom pointed out that 32 of the world’s 50 largest AI companies are based in California and echoed industry complaints that the legislation would hurt innovation.
“Adaptation is essential as we scramble to manage emerging technologies,” Newsom wrote. “This will require a delicate balance. Although well-intentioned, SB 1047 does not take into account that the Al system is used in high-risk areas, involving sensitive decision-making or the use of sensitive data. Instead, the bill applies stricter standards to even the most basic things — as long as the big plan uses them.”
Despite corporate lobbying, many in the AI industry believe legislation is needed. Dozens of AI researchers at leading AI companies called on Newsom in an open letter to sign the bill into law.
“We believe that more powerful AI models may pose greater risks, such as the extended reach of biological weapons and cyberattacks on critical infrastructure,” they wrote. “It is possible and appropriate for borderline AI companies to assess whether the most powerful AI models can cause serious harm, and for these companies to develop appropriate safeguards against such risks.”
The tech workers were joined in their advocacy by some of the biggest names in Hollywood—from JJ Abrams and Ava DuVernay to Mark Hamil and Whoopi Goldberg—who wrote their own open letter in support of SB 1047. Meanwhile, Elon Musk made strange friends with the couch . civil society organizations such as the Electronic Frontier Foundation in calling for its legislation. While Rep. Nancy Pelosi and other influential members of California’s congressional delegation told Newsom they vetoed the legislation, calling it “well-intentioned but ill-informed.”
After Newsom’s vote, Weiner, the bill’s sponsor, said the decision “sets back everyone who believes in corporate governance.”
“The governor’s veto message lists a variety of criticisms of SB 1047: that the bill doesn’t go far enough, but goes too far; that the risk is urgent but we must proceed with caution,” said Wiener in a statement. “SB 1047 was designed by some of the world’s leading AI minds, and any claim that it is not based on concrete evidence is absurd.”
Source link
