OpenAI Nears Deal With Common Sense Media On California Chatbot Safety Measure

OpenAI and the children’s online-safety group Common Sense Media appear close to a negotiated compromise over competing California ballot initiatives aimed at restricting how AI chatbots interact with young people, according to a draft obtained by POLITICO. The revisions reflect talks between OpenAI and Common Sense Media CEO Jim Steyer and would narrow and clarify proposed rules governing minors’ use of chatbot systems.

Key Provisions In The Draft

The draft outlines several major requirements and safeguards:

• Age Verification: AI companies would be required to verify user ages and implement protections for young users.
• Limits On Data Use: The sale of minors’ personal data would be restricted and targeted advertising to children would be banned.
• Parental Controls And Alerts: Platforms must offer parental controls and give parents the option to be notified if a child expresses intent to self-harm.
• Transparency And Audits: Companies would publish child-safety policies and submit to independent audits to confirm compliance.
• Enforcement: The California Attorney General would enforce the rules and levy penalties for violations.

The draft also carves out specific exemptions for certain applications, excluding AI systems used only for commercial business purposes, video-game features that resemble chatbots, and smart speakers from the new requirements.

Context And Implications

The proposed compromise comes as OpenAI and other chatbot developers face increased scrutiny and litigation tied to children’s interactions with AI, including lawsuits alleging harms such as teen suicides. Working with Common Sense Media could help OpenAI blunt political and public criticism while avoiding an expensive, high-profile ballot fight.

Both sides had prepared for a bruising campaign; a signature-gathering drive to qualify an initiative for the ballot can cost more than $10 million. Organizers could begin collecting signatures as soon as next week. The talks may also affect parallel legislative efforts in the California Legislature, where lawmakers recently introduced bills proposing age checks for chatbots and a temporary ban on AI-powered toys for children under 13.

Bottom line: The compromise would not ban children from using chatbots outright but would impose tighter age checks, data limits, transparency requirements, parental controls and outside audits, with enforcement by the state attorney general.