An anonymous reader shared the following report from The Washington Post.
California lawmakers introduced a bill Thursday aimed at requiring companies to test their most powerful artificial intelligence models before releasing them. The groundbreaking proposal could spur regulation across the country as state legislatures increasingly grapple with rapidly evolving technology.
The new bill, introduced by state Sen. Scott Wiener, a Democrat who represents San Francisco, would require companies training new AI models to test their tools for “unsecure” behavior, put in place anti-hacking measures, and use secure methods. This mandates the development of technology in It would be permanently shut down, according to a copy of the bill. AI companies will be required to disclose their testing protocols and what guardrails they have in place to the California Department of Technology. State attorneys general can sue the company if the technology causes “substantial harm.”
Wiener's bill comes amid a surge of state bills addressing artificial intelligence. Policymakers across the country are increasingly alarmed that years of inaction in Congress have created a regulatory vacuum that benefits the tech industry. But California, home to many of the world's biggest technology companies, plays a unique role in setting the precedent for guardrails in the tech industry. “We can't go into software development and ignore what California says and does,” said Lawrence Norden, senior director of elections and government programs at the Brennan Center. He said he believed the bill could be passed by then.
The article points out that there are currently “407 AI-related bills under consideration in 44 U.S. states (according to an analysis by an industry group called BSA the Software Alliance), and several have already been passed.” There is. “The proliferation of state-level bills will increase industry pressure on Congress to pass AI legislation because it may be easier to comply with federal law than with a patchwork of different state laws.”
Even the proposed California law, according to the article, is “largely an evolution of President Biden's October executive order,” which “uses emergency powers to require safety testing of powerful AI systems.” The measure goes further than the executive order, requiring companies to conduct AI-related internal audits and share the results with the federal government. It protects whistleblowers and forces companies to conduct tests. ”
They also added that as the most populous state in the United States, “California has the unique authority to set standards that impact the entire nation.” And while the article notes that groups that supported last year's statement on AI risks helped draft the bill, Weiner also said he consulted with technology stakeholders, CEOs, and activists. “We have done a lot of outreach to our stakeholders over the past year.”
