OpenAI backs California youth AI safety bill

OpenAI's endorsement of SB 1119 builds on a commitment CEO Sam Altman made last year when outlining the company's principles on teens. The company frames the bill as a natural extension of measures it has already implemented, including parental controls, age prediction, and its Teen Safety Blueprint, which are now part of its Model Spec—the rules guiding how its models should behave. These rules prohibit romantic engagement, encouraging emotional dependence, and claiming ChatGPT is human or sentient.

The company emphasizes that teens are the first generation growing up with AI, noting that nearly nine in ten teens who use ChatGPT turn to it for learning or productivity in a given week. It also stresses that the bill recognizes AI is not social media, aiming to preserve access to educational features while adding protections. OpenAI says it looks forward to supporting continued study and collaboration during implementation.

The stakes for OpenAI are partly reputational and partly practical. By supporting this legislation, the company positions itself as a leader in youth safety, potentially influencing how other states and countries approach the issue—though this is inference, not stated in the letter. The outcome hinges on whether Governor Newsom signs the bill, and whether the safeguards prove workable without unduly limiting the learning benefits OpenAI says it wants to preserve.