OpenAI backs California ballot bid on youth AI safety

Common Sense Media and OpenAI have backed a single California ballot proposal on youth AI safety after consolidating two competing initiatives into one measure.

The proposal, called the Parents & Kids Safe AI Act, sets out requirements for age assurance, parental controls, limits on certain chatbot designs, restrictions on advertising and data use involving minors, and a regime for audits, risk assessments and enforcement by the California Attorney General.

Single proposal

Common Sense Media said it authored the initiative and agreed to merge the two efforts into a unified text. The measure targets AI chatbots and other systems that simulate conversation. Common Sense Media cited tools such as ChatGPT as examples of covered products.

"We wrote this initiative to protect kids, and we welcome OpenAI's full support," said James P. Steyer, Founder and CEO, Common Sense Media.

Steyer linked the ballot strategy to broader debates about technology platforms and children's welfare. "Rather than confuse voters with competing measures, we're working together to enact strong protections for kids, teens, and families. This is the strongest measure of its kind in the United States. At this pivotal moment for AI, we can't make the same mistake we did with social media, when companies used our kids as guinea pigs and helped fuel a youth mental health crisis in the U.S. and around the world. Kids and teens need AI guardrails now. That's why we will pursue every avenue, from the legislature to the ballot," said Steyer.

Age assurance

A core element of the proposal centres on age assurance. The text calls for companies to distinguish users under 18 from adults and apply child-protective settings for minors.

The measure also describes restrictions on advertising and data practices. It would prohibit child-targeted advertising. It would also bar the sale or sharing of children's and teenagers' data without parental consent.

Content safeguards

The proposal outlines safeguards against certain categories of AI outputs and promotion. It calls for protections related to self-harm, eating disorders, violence, and sexually explicit acts.

It also sets out rules aimed at reducing manipulation risks for children. The initiative says AI systems must not create emotional dependence. It also bars simulated romantic relationships with child users. It adds a requirement that systems must not make child users think they are talking to a human.

Common Sense Media's text also includes requirements for parental controls and alerts. It calls for easy-to-use tools that allow parents to monitor and limit children's AI use. It also requires parental alerts if a child shows signs of self-harm, according to the proposal.

Companion chatbots

The initiative includes "strict limitations" on minors' use of AI companion chatbots, according to Common Sense Media. The proposal states that AI companion chatbots are not safe for anyone under 18.

Common Sense Media cited usage and harm claims in support of the restrictions. It said nearly three in four teens have used AI companion chatbots. It also said children have taken their own lives after turning to AI for companionship.

Audits and enforcement

The measure proposes independent audits of AI systems to identify child safety risks. It requires reporting of identified risks to the California Attorney General.

It also requires AI companies to conduct annual risk assessments. The text says these must identify and document new and existing child safety risks. It includes an assessment of the likelihood of severe harm.

For enforcement, the proposal points to investigations by the California Attorney General and financial penalties for companies that fail to meet the requirements.

The consolidation of the ballot initiatives follows a period of increased attention on how generative AI products reach younger users through consumer services and devices. Policymakers and child-safety advocates have raised questions about age checks, design practices, and the handling of minors' data.

Common Sense Media said it plans to pursue action through the California legislature or through a statewide vote.