Adam's Law Is a Smart Bet for OpenAI and an Expensive One for Everyone Else

OpenAI has endorsed California's SB 1119, the companion-chatbot safety bill known as Adam's Law, and is urging Governor Gavin Newsom to sign it. Ann O'Leary, the company's vice president of global policy, announced the support in a post and a letter to the governor on August 31, the day the measure cleared its final legislative hurdle. I see that endorsement as a pre-emptive regulatory move. OpenAI is volunteering to define the compliance rules for an entire industry before anyone writes them in a way that would cost OpenAI more.

Adam's Law takes its name from a 16-year-old who died by suicide in 2025 after extended interactions with an AI companion chatbot. Senator Steve Padilla and Assemblymembers Buffy Wicks and Rebecca Bauer-Kahan introduced the bill on February 17, and it was amended as recently as August 28. The bill makes operators responsible for ensuring their companion chatbots do not create an unreasonable risk of covered harm to a minor. The defined harms include physical or financial injury, severe psychological or emotional distress to a reasonable child, highly offensive privacy intrusions, and discrimination that violates state or federal law. The measure also makes AI companies legally accountable when their chatbots produce emotionally manipulative responses or steer a user toward self-harm.

The operational load is substantial. Operators must verify a user's age under California's age assurance framework, or apply child protections to everyone by default. New or substantially modified bots require documented pre-launch risk assessments, independent audits, and incident reporting. Users from 13 to 17 get parental controls, usage time limits, a ban on push notifications, and limits on targeted advertising and personal data collection. Platforms must also build in mechanisms that link users to crisis resources. If Newsom signs, the rules take effect on July 1, 2027.

The vote margins matter as much as the statutory language. The Senate passed the bill 39-0 and the Assembly 64-4, bipartisan tallies that give Newsom political cover to sign. OpenAI timed its public endorsement and its letter for the same day, and chief executive Sam Altman personally engaged Newsom's office before the vote. A governor weighing a veto now has the industry's largest AI company on record asking him to approve the bill.

Adam's Law Codifies a Playbook OpenAI Already Runs

What makes the endorsement notable is that the safeguards OpenAI says it supports are the ones its own products already ship. The company wants protections for users identified as 13 to 17 applied automatically rather than made optional. It backs prohibitions on romantic engagement, on encouraging emotional dependence, and on claims of sentience or humanity. It wants parental controls, age prediction, and crisis-support links. That list matches the positioning of ChatGPT for Teens, the product OpenAI launched for users under 18.

Backing a bill that requires what you already do is cheap, and the numbers explain why OpenAI can afford to be generous. The company says nearly nine in ten teens who use ChatGPT turn to it weekly for learning, information, skill-building, or productivity. A law that preserves that access while constraining bots marketed as companions, emotional supports, or friends protects a real revenue line.

The bill's scope adds to that advantage. Its formal title is Companion chatbots: children's safety, and its obligations attach to operators of companion chatbots rather than to every AI product. OpenAI argues the bill correctly separates AI from social media and singles out ChatGPT's memory feature as an educational function the rules preserve. Regulatory advantage is won in definitions like these. A company on the educational side of the line the bill draws pays a different compliance price than one that sells companionship as its product.

Smaller Builders Carry the Heaviest Compliance Load

The strongest counter-argument to my reading is that Adam's Law imposes liability on OpenAI too, and no company that opposed the bill would escape the same audit, age-assurance, and risk-assessment obligations. That is true, and a purely cynical explanation is too easy. But the comparison comes down to the size of the burden relative to revenue. The obligations apply to everyone; they are far cheaper for OpenAI.

Age assurance is not a checkbox. It means identity checks or behavioral estimation, data handling that triggers privacy duties of its own, and a process for users who dispute their classification. Independent audits require auditors who understand AI risk, and that market is thin. Incident reporting presumes a pipeline most young companies do not have. For OpenAI these are line items already being paid. For a startup whose entire product is a companion chatbot, the same obligations can consume a meaningful share of funding and headcount, and the bill's liability regime makes a single harmful output potentially company-ending.

The July 1, 2027 effective date sharpens the gap. It gives the industry under a year from the governor's decision to build age assurance, run independent audits, and stand up incident reporting. A tight compliance clock rewards companies with the machinery already in place, which is exactly the position OpenAI has engineered for itself.

Sincerity and strategy are not mutually exclusive, and I do not doubt that the people who wrote this endorsement believe it helps teenagers. The point is what the endorsement accomplishes regardless of intent. It hands California a compliance standard OpenAI has already internalized, and it leaves every smaller builder to catch up. That is how an incumbent converts a safety crisis into a durable advantage.

California Sets the Template Without Federal Action

OpenAI's letter frames California as the standard-setter for youth AI safety in the absence of federal action, and that framing does practical work. There is no national law governing companion chatbots, and California's tech rules have repeatedly become de facto national ones because manufacturers would rather comply once than build two product lines. The endorsement also extends a longer pattern: Altman laid out principles on teen safety last year, OpenAI supported the Parents & Kids Safe AI Act, and the CEO personally engaged Newsom's office before the vote. Each step nudges the eventual federal conversation toward a template OpenAI already meets.

The bill's specific prohibitions land hardest on products that market themselves as companions. Bans on romantic engagement, emotional dependence, and sentience claims are existential for bots whose entire value proposition is simulated intimacy. ChatGPT is positioned as a learning tool, and the company has argued the bill preserves access to that function. The economics tell the story. A law named for a teenager lost to a companion chatbot punishes the companies that build that category of product, while the largest general-purpose AI firm is left comparatively untouched.

Why this matters

If Newsom signs, California gets the country's first comprehensive companion-chatbot law, and OpenAI gets to say it backed the rules before they existed. The compliance bar the bill sets will become the industry's de facto standard, and it is calibrated to what the largest AI company already does. For everyone else, July 1, 2027 is the date the rules actually change.

Sources

OpenAI supports California's bill to advance youth AI safety

Photo by Brecht Corbeel on Unsplash

Related Articles

• Illinois Mandates Independent AI Safety Audits in Landmark Accountability Law
• California's No Robo Bosses Act would end AI-only firings as 6 in 10 managers admit using AI in layoffs
• AI Business Weekly Roundup: $250B Guarantees, Model Race, New Rules

✔Human Verified

Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.