All the news about SB 1047, California’s bid to govern AI

Sep 03, 2024 11:23 PM - 6 months ago 222771

California is known for taking connected regulatory issues for illustration data privacy and social media contented moderation, and its latest target is AI. The state’s legislature precocious passed SB 1047, 1 of nan US’s first and astir important frameworks for governing artificial intelligence systems. The measure contains sweeping AI information requirements aimed astatine nan perchance existential risks of “foundation” AI models trained connected immense swaths of human-made and synthetic data.

SB 1047 has proven controversial, drafting disapproval from nan likes of Mozilla (which expressed concern it would harm nan open-source community); OpenAI (which warned it could hamper nan AI industry’s growth); and Rep. Nancy Pelosi (D-CA), who called it “well-intentioned but sick informed.” But peculiarly aft an amendment that softened immoderate provisions, it garnered support from different parties. Anthropic concluded that nan bill’s “benefits apt outweigh its costs,” while erstwhile Google AI lead Geoffrey Hinton called it “a sensible approach” for balancing risks and advancement of nan technology.

Governor Gavin Newsom hasn’t indicated whether he will motion SB 1047, truthful nan bill’s early is hazy. But nan biggest instauration exemplary companies are based successful California, and its transition would impact them all.

  • Adi Robertson

    SB 1047 has passed nan California Senate.

    The Senate was wide expected to walk nan bill, which has now officially cleared each hurdle isolated from a last signature from Governor Gavin Newsom. Newsom has until nan extremity of September to make his call.


    Digital photograph collage of a judge pinch gavel whose hands has excessively galore fingers.

  • Wes Davis

    California legislature passes sweeping AI information bill

    Digital photograph collage of a judge pinch gavel whose hands has excessively galore fingers.

    Illustration by Cath Virginia / The Verge | Photos from Getty Images

    The California State Assembly and Senate have passed nan Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047), 1 of nan first important regulations of artificial intelligence successful nan US.

    The bill, which has been a flashpoint for statement successful Silicon Valley and beyond, would obligate AI companies operating successful California to instrumentality a number of precautions earlier they train a blase instauration model. Those see making it imaginable to quickly and afloat unopen nan exemplary down, ensuring nan exemplary is protected against “unsafe post-training modifications,” and maintaining a testing process to measure whether a exemplary aliases its derivatives is particularly astatine consequence of “causing aliases enabling a captious harm.”

    Read Article >

  • Emilia David

    California legislator files measure prohibiting agencies from moving pinch unethical AI companies

    An image showing a repeating shape of encephalon illustrations

    Illustration: Alex Castro / The Verge

    A 2nd California authorities legislator has introduced bills meant to modulate AI systems, peculiarly those utilized by authorities agencies.

    Senator Steve Padilla, a Democrat, introduced Senate Bills 892 and 893, establishing a nationalist AI assets and creating a “safe and ethical framework” astir AI for nan state. Senate Bill 892 will require California’s Department of Technology to create safety, privacy, and non-discrimination standards astir services utilizing AI. It besides prohibits nan authorities of California from contracting immoderate AI services “unless nan supplier of nan services meets nan established standards.” 

    Read Article >

  • Emilia David

    California lawmaker proposes regularisation of AI models

    AI brains successful a network

    Illustration by Alex Castro / The Verge

    A California lawmaker will record a measure seeking to make generative AI models much transparent and commencement a chat successful nan authorities connected really to modulate nan technology.

    Time reports that California Senator Scott Wiener (D) has drafted a measure requiring “frontier” exemplary systems, usually classified arsenic ample connection models, to meet transparency standards erstwhile they scope supra a definite amount of computing power. Wiener’s measure will besides propose information measures truthful AI systems don’t “fall into nan hands of overseas states” and tries to found a authorities investigation halfway connected AI extracurricular of Big Tech. 

    Read Article >

More