Protecting Kids From AI: Is This Ballot Measure the Right Tool?
Anya Dalal
When OpenAI and Common Sense Media announced the Parents & Kids Safe AI Act, a proposed ballot measure requiring AI companies to filter chatbots for minors, ban the sale of children’s data, and undergo independent safety audits, many viewed it as a turning point for youth technology safety.
The proposal signals that major technology players recognize growing concern about how AI affects children. It promises age estimation, content filters, restrictions on targeted advertising, and mandatory safety audits.
Still, I am concerned that the measure could unintentionally weaken existing protections. The California Initiative for Technology and Democracy and Tech Oversight California have raised issues about how the proposal defines harm and how it interacts with current consumer protection laws. The measure appears to define child protections around “severe harms,” which could leave companies insulated from liability for psychological distress or emotional manipulation by companion chatbots. Exposure to age-inappropriate content may not clearly fall within that threshold either.
My larger concern is structural. Ballot measures in California are difficult to amend once enacted. Meaningful changes often require a two-thirds vote of the California Legislature. Embedding a narrow regulatory framework into law at this stage could limit the state’s ability to strengthen protections as AI technologies and risks evolve. Comprehensive legislation developed through the Legislature would allow more flexibility and refinement over time.
At the same time, I do not dismiss the strategic value of this approach. A ballot initiative creates political leverage. It can push companies to negotiate and may prevent future dilution of youth protections. A constitutional structure can also make it harder for industry pressure to erode safeguards later.
Ultimately, the question is whether this proposal reflects the strongest framework California can build, or whether it locks in a compromise before we fully understand the scope of emerging harms. As signature gathering begins for the November 2026 ballot, Californians will need to weigh the promise of immediate action against the importance of durable, adaptable protections.
Image credit: 1Invites