Tightening Controls on AI Chat Toys for Children! New Laws in New York and California Lead to Regulatory Waves Across Multiple U.S. States

[Ebrun Original] July 8th, as generative AI hardware rapidly iterates and penetrates the consumer market, multiple U.S. states are advancing new regulatory measures targeting children's use of AI companion products.

Among them, New York and California have proposed sales moratoriums or other restrictions on "AI-enabled children's toys with conversational capabilities." States like Washington and Oregon have also begun drafting regulations focusing on transparency, risks of emotional dependency, and minor protection concerning AI companion chat applications.

The S9408A bill currently progressing in New York is considered one of the most aggressive regulatory measures in the U.S. targeting AI chat toys.

This bill aims to prohibit the manufacturing, exchange, distribution, and sale of children's toys with chatbot functionality within New York State, proposing a five-year temporary sales ban to provide time for regulators and researchers to assess associated risks.

According to information released by the New York State Senate, the bill applies to "chatbot toys," primarily targeting AI toy products capable of sustained natural language interaction with children.

The bill's key concerns focus on three main areas.

First, the potential for children to develop excessive emotional dependency on AI toys: as some AI toys can retain long-term user information and respond to children's needs through anthropomorphic language, legislators worry children may blur the line between artificial intelligence and real humans.

Second, issues concerning children's privacy protection, including risks associated with built-in microphones, voice data collection, and cloud data processing.

Third, concerns about the development of children's social skills, questioning whether long-term reliance on AI interaction might negatively impact real-world social relationships.

As of now, the bill has passed review by the state Senate and Assembly, and recently cleared votes in both chambers, proceeding to subsequent processing.

California is also advancing similar measures.

Its state Senate-proposed SB867 bill introduces a four-year moratorium on "companion chatbot toys," restricting the sale of toys with companion chat functionalities to children.

This bill focuses on toys intended for children aged 12 and under, covering manufacturing, sales, and supply to retail channels.

Similar to the New York approach, California's SB867 targets the "sustained interactive capability" of AI toys, not all smart toy products. Ordinary electronic toys, simple voice-controlled devices, etc., do not necessarily fall under its regulatory scope.

The primary policy goal of this bill is to establish a clearer regulatory framework based on further assessment of the psychological and privacy impacts of AI companion products on children.

Beyond targeting physical toys, some U.S. states are expanding their regulatory focus to include software-based AI companion chat applications for children.

Washington State's relevant legislation primarily targets "AI companion chatbots," defined as artificial intelligence systems capable of sustained communication with users via natural language and simulating human-like interactive relationships.

These products differ from traditional customer service bots in their ability to maintain multi-turn conversations, form long-term interactions, and respond to users' emotional needs in an anthropomorphic manner.

Washington's regulatory approach does not directly ban such products but requires companies to assume greater safety responsibilities.

Its regulations mandate that AI companion chatbots must clearly disclose their AI identity to users, avoiding claims of being human. Furthermore, for underage users, companies must implement measures to restrict inappropriate content and prohibit tactics that exploit emotional dependency to prolong user interaction, such as encouraging persistent reliance on AI companionship, fostering feelings of isolation, or simulating intimate relationships.

Additionally, the Washington proposal also focuses on response mechanisms for AI systems when users express crisis situations, including establishing safety measures to identify risk signals and provide relevant support information.

Legislators believe that because AI companion chatbots can simulate emotional exchange, they differ from traditional internet products and thus require additional transparency and safety requirements.

On the West Coast, Oregon is also following the footsteps of California and Washington, promoting minor protection rules for AI chat applications.

Its regulatory direction is relatively close to Washington's, emphasizing AI identity disclosure, restricting simulated emotional dependency, and improving risk response mechanisms, aiming to mitigate potential risks by regulating corporate design and operational practices.

Based on current legislative trends across U.S. states, the regulatory focus is shifting from solely concerning AI-generated content to addressing the long-term interactive relationships formed between AI and users, and the specific hardware products built upon them.

Industry insiders point out that as AI technology evolves from "tool-based applications" to "companion-based applications," the interaction patterns between children and artificial intelligence are becoming a new regulatory topic. In the future, U.S. states may continue to introduce more rules surrounding age verification, data protection, product design liability, and the boundaries of anthropomorphic AI interaction.

Currently, as there is no unified federal regulatory framework for children's AI products in the U.S., states are exploring different regulatory paths through local legislation.

Ebrun will continue to track and report on this development.