LEILA FADEL, HOST:
California took the lead on regulating tech to protect privacy. Now state lawmakers say they expect to set the pace for AI rules. Rachael Myrow reports.
RACHAEL MYROW, BYLINE: By now, you've probably gotten the memo. Large AI models are developing new antibiotics and helping humans communicate with whales but also turbocharging election-season fraud and hiring discrimination.
(SOUNDBITE OF ARCHIVED RECORDING)
SAM ALTMAN: Going off to build a super powerful AI system in secret and then dropping it on the world all at once, I think, would not go well.
MYROW: That's Sam Altman, CEO of OpenAI, the outfit that gave us ChatGPT. That's him almost a year ago now, telling U.S. senators to please pass new laws and create a new agency to force accountability from the big players like OpenAI-backer Microsoft, as well as Amazon, Google and Meta. But there's been little congressional action since, leaving the Federal Trade Commission and other regulators to use knives in a gunfight with laws written long before the advent of AI.
(SOUNDBITE OF ARCHIVED RECORDING)
LINA KHAN: We face basic questions of power and governance.
MYROW: During an FTC summit on AI last month, the agency's chair, Lina Khan, said her staff is alert to the ways AI is already being deployed to abuse market power and take advantage of American consumers.
(SOUNDBITE OF ARCHIVED RECORDING)
KHAN: Will a handful of dominant firms concentrate control over these key tools, locking us into a future of their choosing?
MYROW: But if federal lawmakers are not passing new laws, state lawmakers are taking the lead. California Democratic state Senator Scott Wiener of San Francisco is one of a growing number of lawmakers who's rolled out legislation that could prove a model for other states to follow, if not Capitol Hill.
SCOTT WIENER: I would love to have one unified federal law that effectively addresses AI safety issues. Congress has not passed such a law. Congress has not even come close to passing such a law.
MYROW: Wiener argues his is the most ambitious proposal yet, leaning heavily on the ideas in the Biden administration's October executive order on AI. Wiener's bill would require companies building the largest and most powerful AI models to test for safety before releasing those models to the public. AI companies would have to tell the state about their testing protocols and guardrails, and if the tech causes critical harm, California's attorney general could sue.
WIENER: It's not too late. These large, powerful models that we're talking about in this bill do not yet exist, but they will exist soon, so now is the time to put safety protocols in place to avoid those risks.
MYROW: By one recent count, there are more than 400 AI-related bills pending across 44 states. But when it comes to Silicon Valley tech, not all state legislatures are created equal. Grace Gedye, an AI policy analyst at Consumer Reports, says California lawmakers are poised to lead this legislative discussion because they're better educated than most about what's cooking in AI.
GRACE GEDYE: We can't have a patchwork of state laws, but I think we definitely can't hold our breath because then we could be waiting 10 or 20 years. And in the meantime, consumers will have not the protections they need or the recourse they need to deal with how this technology gets integrated into their lives.
MYROW: Gedye says lawmakers tell her they don't want to make the same mistake with AI as they did with online privacy and social media - that is to say, waiting too long before demanding Silicon Valley protect the rest of us from technology's most harmful effects.
For NPR News, I'm Rachael Myrow in San Francisco.
(SOUNDBITE OF MUSIC) Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.