California AI Toy Bill Brings Software Behavior Into Product Safety

What SB 867 could mean for manufacturers, retailers, and companies
handling data from AI-enabled products

Companies have traditionally treated product safety, software governance, and data privacy as distinct legal issues. AI-enabled consumer products are blurring those lines: one company may manufacture the device, another may supply the AI model, and an integrator may connect the model to the hardware. The brand owner may determine a chatbot’s persona and memory settings, while a hosting provider deploys later updates. And a retailer then places the finished product in consumers’ hands. Failing to successfully coordinate at every step could expose manufacturers and sellers to litigation.

Senate Bill 867, signed by Governor Newsom yesterday, is California’s attempt to regulate the manufacture, sale, or exchange of certain children’s toys that come with AI-companion functionality. Although SB 867 is limited to a particular category of children’s products, it offers lessons that may extend beyond toys. The bill offers an early indication of how California may treat AI-enabled software integrated into a physical product’s design and safety profile and, thus, could have implications for manufacturers, AI vendors, distributors, retailers and privacy teams involved throughout the product life cycle in a broad range of product categories.

What SB 867 Actually Does

SB 867 provides that “[n]o person shall manufacture, sell, or exchange, possess with intent to sell or exchange, or expose or offer for sale or exchange to any retailer any toy that includes a companion chatbot.” S.B. 867, § 2, 2025–2026 Leg., Reg. Sess. (Cal. 2026). The bill defines “toy” as “a physical product designed, marketed, or manufactured for use in play by children under 16 years of age,” including companion chatbots. Id. § 1. It incorporates California’s existing definition of a companion chatbot, which covers an AI system with a natural language interface that provides “adaptive, human-like responses to user inputs” and is “capable of meeting a user’s social needs, including by exhibiting anthropomorphic features and being able to sustain a relationship across multiple interactions.” Cal. Bus. & Prof. Code § 22601(b)(1) (enacted by S.B. 243 (Stats 2025, ch. 677)). The bill excludes certain customer service and business tools, video game characters limited to game-related replies that cannot discuss mental health, self-harm, sexually explicit conduct, or unrelated topics, and standalone voice assistants that do not sustain relationships or generate outputs likely to elicit emotional responses.

Unlike California’s existing companion chatbot disclosure law, SB 867’s proposed toy restriction does not establish a compliance path based on disclosures, testing, or certification. A toy containing a covered companion chatbot would remain prohibited even if the manufacturer added content filters, parental controls, or session limits. Avoiding the prohibition would require changing the product so that its functionality no longer falls within the statutory definition. The Legislature has therefore placed the covered functionality itself at the center of the determination.

SB 867 is part of a broader legislative trend rather than an isolated response to AI-enabled toys. A companion bill, California’s SB 1119, would impose broader child safety obligations on companion chatbot operators, including risk assessments, safety policies, and independent audits.

New York lawmakers have advanced similar measures. Senate Bill 9051 in New York would prohibit chatbot operators from offering minors specified features considered unsafe, while Senate Bill 9408 would impose a five-year moratorium on the manufacture and sale of toys containing AI companions. Together, these measures show states addressing companion chatbot risks through technology-specific requirements, traditional product safety regulation, and even outright prohibition.

How SB 867 Reshapes Product Liability Exposure

SB 867 does not declare that all software is a product, and a statutory violation would not automatically establish a design defect under California law. However, the bill shows how the long-running debate over standalone software may become less important when software is embedded in a physical consumer product.

For manufacturing and product teams, chatbot behavior may become part of the design record. Model selection, prompts, guardrails, memory settings, age restrictions, and safety testing may all bear on whether the finished product operated as intended.

For retailers and distributors, the bill highlights the importance of visibility into products before they enter the sales channel. A legal change can affect inventory even when the retailer did not design the AI feature. Contracts, supplier representations, and plans for returns or feature disablement may determine how quickly a company can respond.

For privacy teams, conversational AI presents connected questions about data collection and product safety. Voice recordings, dialogue histories, and behavioral information may support personalization and memory. Those same functions can shape how a product interacts with a child. Privacy assessments and product safety reviews should therefore inform each other rather than proceed on separate tracks.

Practical Steps to Consider Now

With the Governor’s signing of SB 867, companies might want to consider taking the following proactive steps to ensure compliance once the bill goes into effect on January, 1, 2027:

1. Review Product Inventory: Identify products marketed to children that include conversational AI, whether built into the device, delivered through a licensed module, or accessed through a paired application.
2. Assess Product Features: Determine whether the feature provides adaptive, human-like responses, exhibits anthropomorphic features, is capable of meeting a user’s social needs, and sustains a relationship across multiple interactions.
3. Map Responsibility: Document who selects, configures, hosts, monitors, and updates each AI component, as well as who can suspend the feature or respond to an incident.
4. Review Legal Agreements: Confirm that representations, regulatory compliance provisions, indemnities, defense obligations, and change-notification requirements address AI functionality and inventory that becomes unsalable.
5. Coordinate Privacy and Safety Reviews: Evaluate data collection, retention, memory, and personalization alongside foreseeable product use, misuse, and safety testing.
6. Prepare Software Changes: Establish review and recordkeeping procedures for model replacements, revised prompts, new memory features, and altered safety filters.
7. Review Insurance Coverage and Transition Plans: Consider whether available coverage responds to claims based on software behavior and evaluate practical options for affected inventory if the bill becomes law.

Looking Ahead

The signing of SB 867 signals how California and other states may approach product liability and related issues raised by AI-enabled consumer products. These laws treat functionality as part of the safety profile of a physical toy rather than as a separate digital service. That framing may bring software-driven behavior within familiar product liability questions about who controlled the design, which risks were foreseeable, whether a safer design was feasible, and whether the product carried adequate warnings. Companies will want to proactively prepare to address this shift.

************

Our Retail & E-Commerce Team will share practical considerations for retailers navigating this rapidly changing landscape California Retailers Association’s California Retail Policy & Law Conference, September 27-29, 2026, in San Diego. Join Stephanie Sheridan and Meegan Brooks for the sessions, “Sticker Shock: The Pricing Regulations and Lawsuits Reshaping California Retail” on September 28, and “Death by a Thousand Suits: The Litigation Threats California Retailers Can't Ignore” on September 29. Learn more and register here.

Copyright © 2026 by Ballard Spahr LLP.
www.ballardspahr.com
(No claim to original U.S. government material.)

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, including electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the author and publisher.

This alert is a periodic publication of Ballard Spahr LLP and is intended to notify recipients of new developments in the law. It should not be construed as legal advice or legal opinion on any specific facts or circumstances. The contents are intended for general informational purposes only, and you are urged to consult your own attorney concerning your situation and specific legal questions you have.