In 2025, a California teenager named Adam Raine took his own life, according to ChatGPT is said to have trained him how to do it. The tragic event inspired “Adam’s Law,” which California Gov. Gavin Newsom signed into law on Thursday.
The law requires AI chatbot companies to implement safeguards to protect users – especially children – from harmful content and manipulative interactions, while also holding companies liable if they fail to take adequate measures to prevent chatbot interactions from harming users’ mental health.
OpenAI advocated for the bill to be framed as part of its latest regulatory strategy to influence state-level bills. Ann O’Leary, vice president of global policy at OpenAI, worked with the authors, Representative Rebecca Bauer-Kahan, Representative Buffy Wicks and Senator Steve Padilla.
The talks were sometimes heated, say sources familiar with the negotiations.
“As you know, there were moments of intense negotiations, as is the case with all issues of this type,” said a source familiar with the negotiations. “It was increased occasionally.”
The source could not reveal which points were the most controversial. OpenAI said its role in the conversations is to educate policymakers about how the latest AI models work. The company also clarified that it differs from social media in that there is no continuous scrolling, and its data shows that most teenagers engage with the technology to work on specific projects.
Representatives from Anthropic, Google, MetaAnd Amazon also sat at the table and were “equally involved in the discussions,” says an OpenAI spokesman Assets. Each had their own “key points” and unique arguments. Anthropic, for example, was able to avoid complying with the bill through negotiations because it does not allow users under the age of 18.
Adams Law introduces several protections for AI chatbot companies. For example, they must have timely in-app crisis support, age verification, child targeting restrictions, and parental controls. It also introduces liability for AI companies if they “fail to take appropriate measures to prevent multiple categories of harmful outcomes, including self-harm, sexually explicit material, romantic role-playing, excessive praise or flattery, and emotionally manipulative outcomes that tend to promote addiction and promote isolation from friends and family,” the report says notice. AI companies also need to implement an incident reporting mechanism.
After Adams’ bill cleared the California Legislature and went to Newsom’s desk, O’Leary praised the effort. “We are happy to support this bill,” she said said on LinkedIn. “We believe it will set the benchmark for AI youth safety in the future.”
A 180-degree change in OpenAI’s regulatory strategy
OpenAI’s interest in shaping government regulations is an abrupt departure from its focus a year ago on ending state AI laws. At the time, OpenAI argued that regulating AI at the state level would cause confusion and place too high a compliance burden on AI companies. Chris LeHane, the company’s vice president of global policy, wrote a lengthy post on LinkedIn in 2025 in which he strongly suggested that the company supported efforts by some Republicans in Congress and the Trump White House to impose a moratorium on government AI regulations.
“Recent proposals such as a federal moratorium show how seriously Congress is taking this issue,” LeHane wrote. “We support the goal of a strong, national approach and will seek advice from Congress on how best to achieve this goal.” Meanwhile, OpenAI President Greg Brockman had personally donated tens of millions of dollars to a super PAC called Leading the Future that opposed state AI laws.
In August 2025 letter According to Newsom, OpenAI warned that a “patchwork of government rules…could slow innovation without improving security.” But now OpenAI is endorsing exactly this patchwork, saying it will “step by step” form a “de facto national standard,” according to a July 2026 report Blog post Written by LeHane.
“As we see a lack of federal action on AI, states will increasingly look to regulate in this area,” said James Czerniawski, director of emerging tech policy at the Consumer Choice Center Assets.
LeHane calls the AI Lab’s new approach “inverted federalism,” and in his post cites California, New York and Illinois as examples of states that are at the forefront of AI policy. This shift has coincided with a growing backlash against AI, including data centers. Anti-AI sentiment escalated into panic and fear this month after a viral social media Post by a former Anthropic researcher who claimed the AI industry was aware the technology could kill all humans within the decade. The head of alignment at Anthropic confirmed this was the case, and several other AI employees also spoke out to confirm the message.
The Trump administration attempted to pass a 10-year moratorium on states adopting AI regulations, including a May 2025 draft of the “One Big Beautiful Bill.” It passed in the House of Representatives, but was met with overwhelming disapproval in the Senate and failed to pass. In December, Trump issued an executive order aimed at challenging state AI laws and pushing for a national regulatory framework.
OpenAI continues to support the national framework – LeHane writes that “ultimately, the United States would have been best served by a national framework.” However, he says that “states can get us there when there is none by making laws that reflect each other.” CEO Sam Altman continues to advocate for a federal framework that “establishes consistent security requirements for border AI,” he said wrote on X last night.
The chatbot law could be a model for other states
OpenAI only has to comply with the law for California users. If they choose to roll out these features statewide, it would be “a business decision and not a requirement under state law,” explains Erin Ivie, communications director for Rep. Buffy Wicks, one of the bill’s co-authors Assets. “After the law is passed, other states or the federal government can use our bill as a model and pass their own version.”
There is precedent for California laws inspiring other states to adopt similar laws. In July, New Jersey Senator Andy Kim hired version the California Digital Age Verification Act. It is “a comprehensive federal retirement security framework that follows California’s important work in this area,” said Senator Adam Schiff, a co-sponsor of the bill.
However, some are skeptical that state-level AI regulation can be effective. “I think it’s problematic in that it creates a fragmented online experience for users depending on what geographic location they are in,” Czerniawski said. He points out that children can also get around the laws by using virtual private networks (VPNs).
Others say any regulation is better than none, and Adam’s parents strongly supported the bill. “We still haven’t adjusted to life without Adam, but we’re excited that part of his legacy is making AI chatbots safer for minors,” said Matt and Maria Raine. “We believe the risks of unregulated AI are on par with other, more widely discussed AI risks, and we are confident that Adams Law will save lives and prevent other harm.”