Mattel and OpenAI Face Backlash Over AI Toys

Consumer advocates are warning of the risks posed by the new partnership between Mattel and OpenAI to create AI-powered toys.

Public Citizen co-President Robert Weissman demands Mattel more transparency, and to reveal the details about their upcoming product.

“Endowing toys with human-seeming voices that are able to engage in human-like conversations risks inflicting real damage on children,” Weissman said, as first reported by Ars Technica.

He fears these toys could harm social development, interfere with real-life friendships, and cause long-term psychological harm. “Mattel should not leverage its trust with parents to conduct a reckless social experiment on our children by selling toys that incorporate AI,” Weissman added.

The first product from the partnership will not target children under 13, according to an anonymous Axios source, which Ars Technica suggests is likely due to OpenAI’s age restrictions. Critics argue that the age restriction does not provide sufficient protection.

OpenAI declined to comment, while Mattel has yet to respond to Ars Technica’s inquiry. The first product from the partnership will be announced this year and released in 2026, according to Mattel’s press release, which states that the collaboration will support AI-powered products and experiences, based on Mattel’s brands.

However, critics like tech executive Varundeep Kaur and digital safety expert Adam Dodge warn that AI toys may expose children to privacy breaches, biased content, or confusing chatbot replies. Kaur also flagged the danger of AI hallucinations, saying these toys could give “inappropriate or bizarre responses” that are unsettling for kids.

Ars Technica reports that critics such as Varundeep Kaur and Adam Dodge, who are a tech executive and digital safety expert respectively, express concerns that AI toys could lead to privacy violations, biased content delivery, and confusing chatbot responses.

Kaur highlighted the risk of AI hallucinations, which could cause toys to generate disturbing or strange responses that might unsettle kids. He also added that further risks may be linked to the toysrecording “voice data, behavioral patterns, and personal preferences.”

Ars Technica reports that Dodge added, “unpredictable, sycophantic, and addictive,” and warned of worst-case scenarios, like toys promoting self-harm. Both experts called for strict parental controls, transparency, and independent audits before any launch.

Indeed, researchers from MIT have issued a separate but related warning about the addictive nature of AI companions.

Mattel has faced similar backlash before. In 2015, the company released “Hello Barbie,” a Wi-Fi-connected doll that listened to kids and responded using cloud-based AI, as reported by Forbes.

Critics at the time, including cybersecurity expert Joseph Steinberg, warned that the toy posed a massive privacy threat. Hello Barbie recorded and uploaded children’s conversations to a server operated by a third party, ToyTalk, which shared the data with vendors to improve AI systems.

Steinberg pointed out that children often confide deeply personal thoughts to their dolls—sometimes discussing fears, family issues, or school problems. “Would you want recordings of their intimate childhood conversations to persist in the hands of unknown parties?” he asked, as reported by Forbes.

Privacy experts argue that unless companies offer plain-language warnings on packaging, many parents will unknowingly expose their children’s private lives to corporations under the guise of convenience and entertainment.