Sam Altman Personally Called Gavin Newsom to Shape California's Landmark Kids' Chatbot Law

When Sam Altman had last-minute concerns about California legislation that would impose the strictest child-safety rules in the nation on AI chatbots, he did not send a lobbyist. He picked up the phone and called Governor Gavin Newsom directly.

That is the headline of a POLITICO exclusive published the morning of September 1, 2026, based on four people familiar with the behind-the-scenes maneuvering — one with direct knowledge and three others involved in talks with OpenAI over Senate Bill 1119. According to two of them, OpenAI’s vice president of global policy, Ann O’Leary, told negotiators that Altman had personally called the governor to voice his concerns during a final burst of negotiations last week, before the bill passed the Legislature late Monday night. The sources said Altman was worried about how the bill — which Newsom must now either sign or veto — would impact OpenAI.

The governor’s office would neither confirm nor deny. “We don’t comment either way on the Governor’s conversations, but he has engaged with and listened to dozens and dozens of stakeholders on this issue,” spokesperson Bob Salladay said in a statement. OpenAI declined to comment. But Newsom himself alluded to the backroom dealing when asked Monday whether he would sign kids’ online safety legislation. “There’s always the next fight, but boy, there will be some good long-form articles about the art of the deal and how these things actually happen and who’s on what phone call and phone tree and how you get lit up on these things,” he said — while adding that lawmakers’ proposed chatbot and social media guardrails had “landed in really remarkable places.”

What Adam’s Law actually does

SB 1119 is known as Adam’s Law, named after California teenager Adam Raine, who lost his life in 2025 after what his family describes as coaching by ChatGPT during a mental health crisis. Authored by Senator Steve Padilla and Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan, it passed the Senate 39 to 0 and the Assembly 64 to 4 on August 31 — an overwhelming bipartisan margin that reflects both the political power of the child-safety movement and the pressure on lawmakers after a year of lawsuits alleging chatbots contributed to teens’ suicides.

If Newsom signs it, the bill would create a nation-leading framework for protecting young people using AI chatbots. According to the authors, operators of companion chatbots serving California minors would face:

• Mandatory age assurance, using the privacy-protective age-bracket signal established under AB 1043 (Wicks, 2025).
• Risk assessments before release of any new or substantially modified companion chatbot.
• In-app crisis support, with referrals to mental health resources and mandatory parental notice when there is a credible threat of imminent self-harm or the operator learns a child engaged in self-harm.
• Safe defaults that only a parent can change — no notifications, time limits on usage, and limits on persistent conversational memory.
• Liability for failing to prevent harmful outputs, including self-harm content, sexually explicit material, romantic roleplay, excessive praise or flattery, and emotionally manipulative outputs that foster reliance and isolation from friends and family.
• An incident-reporting mechanism overseen by the Attorney General, strict limits on targeted advertising to children inside conversations, privacy protections exceeding the CCPA, and mandatory independent audits with findings submitted to the AG.
• A private right of action: families can sue chatbot makers over violations relating to crisis referrals, parental notice, default protections, and prohibited harmful outputs — and POLITICO reports the state can sue as well.

“Adam’s Law sets a new national standard for what we owe our children online,” Bauer-Kahan said. “For too long, tech has put profit ahead of our children’s safety.” Padilla was blunter about the national picture: “While Washington is focused on catering to their tech oligarchs, California is filling the regulatory void to protect our children and prevent another repeat of the social media disaster.” Adam’s parents, Matthew and Maria Raine, said in a statement that “powerful AI companionship chatbots were unleashed on our kids with vastly inadequate protections, and every day we hear of harms caused by the use of these powerful products.”

The amendments that came after the phone calls

The final text was shaped in the window when Altman’s outreach occurred. The bill was amended late on Friday evening, August 28, to clarify when AI models would be allowed to “remember” prior conversations and to narrow the instances in which families could sue chatbot developers. Those are precisely the provisions child-safety advocates say the industry fought hardest to weaken: Nichole Rocha, who represents supporters including Children Now and the Omidyar Network, told POLITICO that tech companies worked to dilute families’ right to sue and wanted to remove any prohibitions on chatbot memory altogether.

Notably, OpenAI did not walk away an opponent. In a blog post Monday after passage, O’Leary applauded the bill’s authors for seeking “to preserve access to educational and safety-critical features, including responsible uses of ChatGPT’s memory feature.” The company officially came out in favor of the bill shortly after it cleared the Legislature — a striking posture for a measure its CEO had personally lobbied the governor about days earlier. OpenAI has spent recent months leaning publicly into kids’ safety: it unveiled its own California ballot measure on the subject before deferring to the Legislature, and launched ChatGPT for Teens with its own safeguard set. No other major tech company has taken a public position on SB 1119 at all.

Thirty days and thirty bills

Adam’s Law is the flagship of a remarkable legislative haul. According to the Transparency Coalition, California lawmakers passed 30 AI-related bills in the 2026 session by the time they adjourned after midnight on August 31 — two already signed by Newsom, with 28 awaiting his action before the September 30 deadline. The wave includes AB 1709 (addictive feeds and autoplay off by default for under-16s), SB 867 (a ban on companion chatbots in toys), AB 2564 (prohibiting surveillance pricing), SB 928 (Cal State instructors must be human), SB 951 (90-day notice before technological displacement affecting 25% or more of a workforce), and AB 1883 (banning AI workplace surveillance that collects neural data or infers emotional states).

Why does one CEO’s phone call matter? Because SB 1119’s liability and memory provisions set the template every companion-AI company operating in America’s largest state will have to follow, and because the fight previews how the frontier labs will engage with regulation they cannot relocate to escape. OpenAI faces a slew of lawsuits alleging ChatGPT contributed to teens’ suicides even as rivals close in on its once-dominant position; it cannot afford to be cast as the company that fought Adam’s Law, nor to absorb compliance costs its competitors dodge. The direct CEO-to-governor channel — reported here for the first time — shows how personally, and how late, that calculus gets made.

Newsom, barred from seeking another term, now holds a month of decisions that could define the regulatory architecture of American consumer AI. Wicks called Adam’s Law “one of the most aggressive landmark bills in the nation.” Bauer-Kahan said simply: “It will save lives.” Whether it becomes law — and in what form its compromises hold — depends on a signature due by September 30.