The Kids Might Be All Right

Two months ago, bipartisan outrage across the country persuaded the United States Senate to defeat a national moratorium on state AI laws and protections by a vote of 99-1. Washington hasn’t passed significant tech legislation for decades. Big Tech tried to block state action for a decade more. But governors, attorneys general, state legislators, and parents made clear that the lives of kids and teens are at risk right now.

This past week the California legislature showed why that victory was so important. By strong majorities, state legislators passed four measures to save kids’ lives and make AI safer that now just need Governor Gavin Newsom’s signature to become law. With these bills, Sacramento sent a clear message: don’t let tech companies write all their own rules for kids.

New Limits on Teen AI Companions

No issue cries out for more urgent action than the rising teen use of AI companions. In April, Common Sense Media released risk assessments showing that social AI companions pose unacceptable risks to children and teens. Last year, a 14-year-old shot himself to death during a conversation with a Character.AI chatbot. This spring, a 16-year-old hanged himself after prolonged discussions of suicide with ChatGPT.

That’s why California Assemblymember Rebecca Bauer-Kahan introduced and Common Sense Media sponsored the Leading Ethical AI Development (LEAD) for Kids Act, which will prohibit companies from offering emotionally manipulative AI chatbots that encourage kids and teens to engage in self-harm, violence, use of drugs or alcohol, disordered eating, or thoughts of suicide. Tech industry trade groups lobbied for much weaker protections. But Bauer-Kahan stood her ground: “We are not experimenting on California’s children. It is not the right thing to do, it’s not safe and this bill will stop that.” Forty state attorney generals, including Rob Bonta of California, sent the same message, warning companies: “We wish you all success in the race for AI dominance. But we are paying attention. If you knowingly harm kids, you will answer for it.” The LEAD for Kids Act passed the State Senate 31-6 and the Assembly 60-8.

Social Media Warning Labels

The more powerful AI becomes, the more addictive the social media apps it powers become as well. Two years ago, Surgeon General Vivek Murthy issued a social media and youth mental health advisory declaring that “social media can have a profound risk of harm to the mental health and well-being of children and adolescents.” Last year he proposed warning labels on social media platforms to let parents and teens know they’re a hazard to young people’s mental health.

Forty-two state attorneys general endorsed social media warning labels. Once again, industry fought the idea, denying that social media played a role in the youth mental health crisis. One tech industry group declared that “slapping a warning label on social media is like a broken fire alarm going off with no evidence of smoke.” But legislators – many of them parents – know better. The Minnesota and New York legislatures passed social media warning label legislation earlier this year. This past week, the California legislature did the same, passing legislation authored by Assemblymembers Bauer-Kahan and Buffy Wicks and sponsored by Attorney General Bonta and Common Sense Media.

Age Assurance

Parents have long sought a safe, privacy-protecting way to keep kids and teens from using apps and websites that aren’t age-appropriate. Tech companies had the data and the means to do so but couldn’t agree on a solution because app stores and operating systems each wanted the other to be the gatekeeper. Privacy advocates worried that companies would force users to share private data like drivers licenses every time they logged onto a site.

Over the past year, Assemblymember Wicks negotiated a compromise that all sides could support. Her bill requires app stores and operating systems to determine a child’s age range – under 13, 13-15, 16-17, 18-and-older – and provide an age assurance signal that doesn’t collect or reveal unnecessary personal information. Tech giants and many privacy advocates welcomed the breakthrough, which passed unanimously in both houses.

AI Transparency and Safety

America can only win the AI race and realize AI’s enormous promise if Americans know and trust that it’s safe. Early Saturday morning, after an 18-month battle, California passed Senator Scott Wiener’s landmark AI transparency bill that puts in place strong whistleblower protections and safety disclosure requirements for AI companies that develop powerful frontier models. Last year, tech lobbyists fought a previous version of the bill so vehemently that Governor Newsom vetoed it. This year, Senator Wiener built a carefully targeted bill around the recommendations of a working group of AI experts the governor appointed. It requires the leading AI labs to develop, make public, and follow safety frameworks and risk assessments to prevent catastrophic risk and disclose critical safety incidents.

Far from imposing new hurdles, the bill codifies key safety measures that frontier AI companies have already voluntarily agreed to honor. As one leading company, Anthropic, said in endorsing the bill, “Transparency requirements will have an important impact on frontier AI safety. . . . The question isn’t whether we need AI governance – it’s whether we’ll develop it thoughtfully today or reactively tomorrow.” The Wiener bill passed overwhelmingly in the final hours of the legislative session and, like the other three measures, is now on the governor’s desk.

Much more will need to be done to protect kids and teens from new harms and make up for a lost decade in which America’s leaders failed to act. But California’s landmark actions this past week are a big step forward. We don’t have to choose between raising healthy kids and winning the AI race. America can and must do both.