OpenAI backs California's teen AI bill and launches ChatGPT for Teens

While most adults are still using AI to polish work emails, teenagers are already treating it like an on-demand personal tutor.

According to OpenAI, nearly nine out of ten teens who use ChatGPT each week rely on it for learning, research, or productivity. But as the first generation grows up alongside chatbots, the guardrails have lagged behind the tech.

OpenAI Global Policy VP Ann O'Leary publicly endorsed California Senate Bill 1119, formally urging Governor Gavin Newsom to sign the youth safety legislation into law. With federal rules nowhere in sight, California is once again poised to set the national playbook.

ChatGPT locks down the kids

The bill forces AI developers to verify user ages, undergo independent safety audits, and block harmful interactions like self-harm and sexual exploitation. It also requires platforms to hand parents monitoring tools, connect at-risk teens to crisis support, and restrict targeted ads.

Alongside the policy push, OpenAI rolled out ChatGPT for Teens, an experience built specifically for users aged 13 through 17. The system automatically routes users into this mode if they state they are underage or if internal signals estimate they are under 18.

Critically, these safeguards are baked into the baseline setup. Teens cannot toggle them off in settings or bypass them with simple workarounds.

To keep the focus on education, the mode includes interactive quizzes, visualizations, and a dedicated Study mode designed with teachers to guide students step by step rather than doing their homework for them. The platform also updated its Model Spec to strictly ban romantic roleplay, emotional codependency, and any claims of being human.

Why it matters

For families, this is a massive relief. Instead of an open-ended chatbot that might hallucinate bad advice or encourage unhealthy attachment, teens get a guided learning companion with safety nets parents can actually manage. The trade-off is friction: legitimate academic research into heavy topics might occasionally trigger false alarms.

For the AI landscape, OpenAI is pulling off a classic regulatory flex. By championing a tough bill in California, the company effectively raises the barrier to entry for smaller competitors who cannot afford external audits and specialized safety engineering.

We think OpenAI's most calculated move here is drawing a firm line between AI and social media. Giants like Meta and ByteDance spent years getting battered on Capitol Hill for teenage addiction, and OpenAI is desperate to avoid that same regulatory crosshairs. Framing ChatGPT as an educational utility rather than an attention-trap app is brilliant politics.

Still, serious friction remains. Algorithmic age prediction is notoriously imperfect, and leveraging long-term chat memory to spot concerning behavioral patterns inevitably bumps up against teen privacy expectations.