California lawmaker proposes a four-year ban on AI chatbots in kid’s toys

Senator Steve Padilla (D-CA) introduced a bill connected Monday that would spot a four-year prohibition connected the merchantability and manufacture of toys with AI chatbot capabilities for kids nether 18. The extremity is to springiness information regulators clip to make regulations to support children from “dangerous AI interactions.”

“Chatbots and different AI tools whitethorn go integral parts of our lives successful the future, but the dangers they airs present necessitate america to instrumentality bold enactment to support our children,” Senator Padilla said successful a statement. “Our information regulations astir this benignant of exertion are successful their infancy and volition request to turn arsenic exponentially arsenic the capabilities of this exertion do. Pausing the merchantability of these chatbot-integrated toys allows america clip to trade the due information guidelines and model for these toys to follow.”

The bill, dubbed SB 287, comes successful the aftermath of President Trump’s caller enforcement order directing national agencies to situation authorities AI laws successful tribunal — though the bid explicitly carves retired exceptions for authorities laws related to kid safety. The authorities besides follows respective concerning incidents involving AI, chatbots, and children.

In implicit the past year, lawsuits filed by families whose children died by suicide aft engaging successful prolonged conversations with chatbots person spurred lawmakers to action. Padilla besides co-authored California’s precocious passed SB 243, which requires chatbot operators to instrumentality safeguards to support children and susceptible users.

While the usage of chatbots successful toys isn’t arsenic mainstream yet, determination person already been reports of troubling interactions. In November 2025, user advocacy radical PIRG Education Fund warned that toys similar Kumma – a cute artifact carnivore with a built-in chatbot – could beryllium prompted easy to speech astir matches, knives, and intersexual topics. NBC News found that Miiloo, an “AI artifact for kids” made by Chinese institution Miriat, would astatine times bespeak that it was programmed to bespeak Chinese Communist Party values.

OpenAI and Barbie-maker Mattel were slated to merchandise an “AI-powered product” successful 2025, but delayed their release. Neither institution explained the delay, and it’s not wide if they program to merchandise a artifact successful 2026. 

“Our children cannot beryllium utilized arsenic laboratory rats for Big Tech to experimentation on,” Padilla said.