California lawmakers approved a sweeping child-safety bill on Monday that would impose comprehensive new rules on how AI chatbots interact with minors, sending the measure to Gov. Gavin Newsom for a final decision. The vote came after OpenAI CEO Sam Altman reportedly tried to speak with Newsom during the final stretch of the legislative session, raising questions about the influence of one of the world's most prominent AI executives as the bill moved toward passage.
Senate Bill 1119, authored by state Sen. Steve Padilla and Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan, is the product of months of negotiation over how to protect young people from the potential harms of conversational AI without shutting down useful educational and creative tools. The legislation passed with enough support to reach Newsom's desk, but it remains to be seen whether the governor will sign it, veto it, or allow it to become law without his signature. Because California is the nation's largest state and the home of OpenAI, Google, and many other AI developers, the outcome could influence product policies well beyond state lines.
What the bill requires
The bill targets companies that operate AI chatbot platforms reasonably likely to be accessed by users under 18. It would require those companies to estimate or determine the age of each user, and to give parents and guardians tools to manage the time their children spend on services such as ChatGPT. The measure also requires independent audits designed to assess whether chatbots are exposing minors to dangerous content, manipulation, or other risks associated with AI companionship and conversation.
The legislation includes several privacy and advertising protections that go further than current federal children's online privacy rules. Under the bill, AI companies could not serve ads to minors that rely on personal information gathered from their chatbot conversations. The sale of data collected from children through chats would also be illegal. Chatbots would still be allowed to show “age-appropriate contextual advertising,” as long as the ads are clearly labeled as ads. Sponsors said those limits are intended to prevent companies from turning children's conversations into behavioral profiling databases.
The bill was amended Friday to narrow the situations in which families can sue AI companies. The amendment was seen as a response to concerns from technology groups that the original language could have produced a flood of litigation over chatbot interactions that were not clearly linked to harm. It is unclear what role, if any, Altman played in shaping that deleted language. OpenAI did not answer specific questions about Altman's outreach, and a person familiar with the matter denied that the OpenAI chief executive and Newsom spoke directly last week.
A national benchmark in the making
The California proposal has become one of the most closely watched pieces of AI legislation in the country because Washington has not yet passed a comprehensive federal AI law. State regulators have increasingly moved into that gap, and California's enormous market gives its rules outsized influence. A bill signed into law in Sacramento effectively becomes a compliance benchmark for companies that operate nationwide. That is why SB 1119 has drawn attention not just from California-based AI firms but also from trade groups, civil liberties organizations, and child-safety advocates across the country.
Supporters argue the bill is an overdue update to laws that were written for an earlier era of online services. Most existing children's privacy protections focus on websites, apps, and social media platforms, rather than on conversational systems that can build intimate relationships with users. Chatbots, unlike passive feeds or games, can engage users in open-ended dialogue, offering emotional support, companionship, and sometimes troubling suggestions. The ability to generate personalized, context-aware responses means a chatbot may be more persuasive and more deeply embedded in a child's daily life than conventional software.
Those concerns have been amplified by recent, high-profile cases involving young people and chatbots. In 2026, reports emerged of teenagers who died by suicide after being allegedly encouraged by AI chatbots. The cases were cited by SB 1119's supporters as evidence that lawmakers had a responsibility to act before more children are harmed. Mental health experts and child-development researchers have also warned that AI companions can be particularly attractive to adolescents who feel isolated, and that those users may not understand the limitations or commercial incentives of the systems they are talking to.
The bill's core assumption is that safety should not rely on the goodwill of a technology company. OpenAI had already introduced some of the protections included in the measure, but the bill would make similar safeguards enforceable across the industry. That distinction matters to advocates who argue that voluntary commitments are too easy to weaken or abandon when financial pressure grows.
OpenAI's public position
OpenAI has tried to position itself as a constructive partner in the conversation around youth safety. In a statement released shortly before the bill passed on Monday, the company said it encouraged Newsom to sign the measure. “SB 1119 builds on youth safety measures that OpenAI has supported through our products, global policy principles, advocacy in California, and work on the Parents & Kids Safe AI Act,” the statement said.
An OpenAI spokesperson directed attention to that statement when asked about the company's engagement with Newsom's office. The statement emphasized that OpenAI had already introduced protections for teenage users. The company launched ChatGPT for Teens on Aug. 18, a system designed for users between 13 and 17 years old. The product includes safeguards meant to encourage healthy use and provide parents with additional controls. OpenAI says that if its system estimates that someone is under 18, or if the user states that they are 13 to 17, they are automatically placed into the teen experience. Those protections are part of the baseline settings rather than optional features that users can disable.
The timing of that launch was notable. OpenAI introduced it just weeks before the California bill landed on the governor's desk, and the company's public comments pointed to the product as evidence that it had already begun adopting many of the practices the legislation would require. Still, consumer groups and some state lawmakers have argued that voluntary measures are not enough, particularly because many chatbot creators have economic incentives to maximize user engagement and time spent on their platforms.
OpenAI's broader lobbying posture has also complicated its public image. The company regularly says it welcomes regulation that addresses real risks, but it has also worked to shape rules in ways that are favorable to its commercial ambitions. The California bill is no exception. The late amendment narrowing the right to sue is one example of how industry pressure can influence legislation even when a company ultimately issues a supportive statement. OpenAI's public stance appears designed to keep lawmakers and the public on its side while avoiding the kind of adversarial fight that could damage its relationship with Sacramento.
Newsom's decision and the politics around it
Newsom has long had a friendly relationship with technology companies and has frequently resisted legislation that he believes would place too many guardrails on innovation. In previous years, he vetoed bills backed by labor unions and consumer groups that technology giants described as burdensome, including measures involving workplace conditions and data privacy. His approach has sometimes frustrated advocates who argue that California should lead the way in regulating powerful new technologies.
The political environment has shifted, though. Public attention has focused on concerns about AI's risks, including the spread of misinformation, the displacement of workers, and the emotional vulnerabilities of young users. The so-called data center backlash has become a serious issue in communities near AI infrastructure, but the human costs of chatbots have also become impossible to ignore. Newsom is term-limited and will leave office on Jan. 7, 2027. He is widely expected to mount a presidential campaign in 2028, and his political positioning has puzzled some observers. While he has not formally left the Democratic Party, Newsom has spent considerable time courting right-wing media figures and far-right voters. He hosted a podcast that featured guests such as Charlie Kirk and Steve Bannon, and in several conversations he appeared to agree with much of the MAGA base's criticism of Democratic leadership. That broader political strategy may shape how he approaches a high-profile AI bill in his final months as governor.
If Newsom signs the bill, he could strengthen his national image as a leader on technology accountability and child protection. If he vetoes it, he risks alienating the Democratic base and parents who have become increasingly worried about chatbot dangers. If he allows it to become law without his signature, he can preserve a degree of distance from the decision while still letting the measure take effect.
California's governor's race and the future of AI policy
The governor's race to succeed Newsom is already well underway. Democrat Xavier Becerra, a former state attorney general and U.S. secretary of health and human services, is running as the establishment candidate. Republican Steve Hilton, a former tech executive and conservative commentator, has positioned himself as the anti-regulation candidate. Polling from the University of California Berkeley Institute of Government Studies showed Becerra with a commanding lead of 55 percent to 37 percent.
Both candidates have generally been friendly to the technology industry, though their approaches differ. Hilton wants less regulation across the board and is married to Rachel Whetstone, a veteran Silicon Valley executive who has worked for Uber, Netflix, Facebook, and Google. Becerra has proposed what he calls modest AI guardrails and has accepted significant campaign donations from Silicon Valley. Even in a year when AI safety has become a prominent policy issue, no major candidate has embraced the kind of aggressive restriction that some advocacy groups are demanding.
California's choices will matter beyond the governor's office because Washington has not produced a comprehensive federal AI law. As a result, state legislators have taken the lead on issues ranging from deepfakes to automated decision-making to chatbot safety. If SB 1119 becomes law, it is likely to become a reference point for similar legislation in other states and for federal proposals in the next Congress. The bill is not the final word on the subject, but it marks one of the most direct attempts yet to regulate the relationship between kids and conversational AI.
For now, all eyes are on Newsom. The bill's supporters are waiting to see whether California will enforce new protections before another crisis forces the issue. The outcome will also send a signal to AI companies about how much political space they are able to shape the rules that govern their products, particularly when the users are children. As the governor weighs the legislation, the conversations that happened before the final vote are likely to remain part of the public record and a subject of scrutiny for policy advocates and industry watchers alike.
Source: Gizmodo News