California’s New AI Toy Law Redefines Software as a Physical Safety Hazard
moved to ban companion chatbots built directly into children's toys, signing Senate Bill 867 into law and framing software behaviour as part of a physical product's safety profile rather than a…

moved to ban companion chatbots built directly into children's toys, signing Senate Bill 867 into law and framing software behaviour as part of a physical product's safety profile rather than a separate digital service, according to reporting by London Insider.
The Product, Not the Patch
Under SB 867, no manufacturer can sell or exchange a toy that includes a companion chatbot — and there is no compliance path through disclosure, content filters, parental controls, or session time limits. The only escape is redesigning the product so its functionality no longer meets the statute's definition of a covered companion chatbot.
For anyone who evaluates toys through the lens of material integrity, this is a striking distinction. Regulators are no longer asking whether a company wrapped a risky feature in enough safeguards; they are asking whether the feature belongs in a child's hand at all. That is the same standard we apply when evaluating finishes, fasteners, and fibre content in a wooden toy: the question is not "how safely did you seal the hazard" but "is the hazard in the product to begin with."
Why a Wooden Toy Maker Should Care
The bill's reasoning extends well beyond chatbot toys. It sits inside a broader legislative push — including New York's Senate Bill 9051, which would prohibit chatbot operators from offering minors certain features deemed unsafe, and Senate Bill 9408, proposing a five-year moratorium on AI companion toys — toward treating software-driven behaviour as part of a product's design-defect analysis under California law.
That framing matters even for brands whose catalogues contain zero electronics. SB 867 lays bare how fragmented AI product supply chains have become: one firm builds the device, another supplies the model, an integrator connects them, the brand owner scripts the persona and memory settings, a hosting provider pushes later software updates, and a retailer ships the finished box. When something goes wrong, liability flows across every link. Wooden toy workshops watching this unfold should note how the law refuses to let any single party hide behind the next.
For small studios and independent makers, the practical lesson is structural. Keep the supply chain legible. Know where every board, bead, and finish originates, document the chain from forest to playroom, and treat traceability the way California now treats chatbot software — as a non-negotiable layer of the product's safety profile, not an afterthought.
What to Watch
SB 867 does not declare all software a "product" under California law, and a statutory violation will not automatically establish a design defect in court. But the regulatory tone has clearly shifted. Similar bills are surfacing across U.S. states, and companies across categories — not just toy manufacturers — are being told to prepare for the change.
For families and educators investing in developmental play, the takeaway is straightforward. The toys that will hold up under the toughest scrutiny are the ones whose value lives in the material itself — grain, weight, joinery, finish — not in a server update pushed overnight. Tactile integrity, once a design preference, is becoming a regulatory moat.