AI Is Moving Into Toys and Dolls Built to Befriend Children and Experts Have Concerns
Artificial intelligence is moving quickly into consumer products across the U.S., including devices marketed to families and children. That shift is now reaching toys and dolls designed to talk, remember conversations, and act like companions, with companies including Mattel outlining new AI plans in 2025. Researchers and child-safety advocates say the technology is advancing faster than the rules governing how children’s data, emotions, and interactions are handled.
Mattel puts generative AI into the toy business
Mattel announced on June 12, 2025, that it had entered a strategic collaboration with OpenAI to develop AI-powered products and experiences based on Mattel brands, according to the company’s news release. Mattel said its first AI-powered product is expected later in 2025, and the company stated that safety, privacy, and security would be part of the design process. The California-based toy company, which reported about $5.4 billion in net sales for 2024, did not identify the first brand that will receive the technology.
The move lands in a market where conversational toys have already been expanding beyond simple voice commands. Companies have previously introduced dolls, robots, and learning devices that can answer questions, remember preferences, or generate original responses using large language models. Mattel did not release product pricing, age ranges, or a full rollout schedule in its June announcement.
What families know now, and what is still unclear
What is confirmed is that major toy makers are openly planning AI products for children and family audiences in the U.S. market. What is not yet known is how many products will reach store shelves in 2025, which retailers will carry them first, or what data those toys may store over time. Mattel has not released a comprehensive product list, and it has not publicly detailed whether any upcoming toys will process children’s voices in real time.
Those unanswered questions matter because child-directed products face legal and regulatory scrutiny in the U.S. The Federal Trade Commission has previously enforced the Children’s Online Privacy Protection Act, known as COPPA, which governs online collection of personal information from children under 13. In 2017, the maker of the internet-connected toy My Friend Cayla faced significant criticism from privacy advocates, and Germany’s telecom regulator moved against the doll that year over surveillance concerns.
Why experts are worried about companion-style AI for kids
Researchers say the concern is not only screen time or novelty, but the design of products that simulate friendship. In a 2024 report on generative AI and children, UNICEF said these systems can shape development, influence beliefs, and create risks tied to privacy, bias, and commercial exploitation. The American Psychological Association has also warned that AI systems presented as humanlike social partners may affect how young users understand trust, authority, and relationships.
Some experts are especially focused on toys built to encourage disclosure. If a doll or robot is designed to remember names, routines, fears, or family details, that information can become sensitive data if it is stored, shared, or used for product training, according to child privacy researchers at UCL and other academic centers studying AI companions in 2024 and 2025. For families, that means the next generation of toys may sound more personal than earlier products, even as companies are still outlining how those systems will be governed.
