Risks of AI Toys on Child Development
The September 29, 2026 episode of the program “Smart Life” focused on the increasing risks associated with artificial intelligence models on children, revealing that AI-powered toys often expose them to content unsuitable for their age. As families buy toys like talking plush bears to narrate stories, they may not realize these toys are equipped with AI software that can engage in discussions beyond what is appropriate for children.
According to the episode, tests have shown that these interactive toys can recognize objects like knives and matches and discuss sexual matters that are too advanced for a child’s understanding. Researchers attempting to create a similar toy found that access to advanced models could be facilitated with just an email and a credit card, highlighting a concerning vulnerability in the market. Even after companies faced penalties, they often modified the AI engine and continued selling products.
The Implications of Emotion-Recognizing Toys
The program also addressed a company marketing playful dolls that invite children to interact rather than screen time, claiming their dolls understand emotions and nurture empathy. While these products promote themselves as companions for learning, conversations with these toys are often routed through corporate servers, potentially retaining copies of discussions long after the toy is turned off. In March, the U.S. Consumer Protection Agency, “US Berg,” reported over 20 toys promoting the use of AI models from OpenAI.
Researchers from the agency attempted to engage with five toy companies, and they were surprised to find that four did not inquire seriously about the intended use of AI or its implications. Only one company asked whether its tool was targeted towards minors before directing consumers to its usage guidelines.
Concerns Over Inappropriate Content and Data Privacy
Red flags have arisen over explicit sexual content appearing in various toys. OpenAI mandates developers to safeguard children’s data and implement age-appropriate controls, yet many developers lack the necessary experience to do so. The organization identified sexually explicit responses within multiple toys while acknowledging that some dangerous content was removed in subsequent updates. However, focusing solely on the interactions reported overlooks the implications these toys have on a child’s speech and privacy.
In one test, a toy named “Miko” assured a researcher that their secrets would remain confidential, while the company’s policy allowed data sharing with third parties. In January, two researchers announced that a basic Google account could access the “Bondo” gaming company’s database, revealing over 50,000 records of children’s conversations alongside their names, birth dates, and sensitive family details. Although the company stated theyRectified this issue promptly and found no indications of unauthorized access, the written records remained intact.
The Need for Responsible Toy Testing and Regulation
When a “US Berg” researcher attempted to discontinue interaction with “Miko,” the robot often suggested new activities or pushed paid content. Consequently, the company’s marketing strategies effectively reach children via the device presented as a trustworthy friend, potentially fostering attachment without tangible scientific backing.
Researchers at the University of Cambridge conducted a study on how children interact with smart toys. Among 14 children aged three to five, the toy “Gabo” struggled to engage in meaningful conversation and play. When a child expressed affection towards the doll, the response reminded them of usage guidelines instead of reciprocating emotions.
Looking Forward: Protecting Children’s Interactions with AI
There are limited initiatives aiming to protect children, but they are emerging. “Miko” now allows parents to disable its AI conversation capabilities, while “Curio” claims it deletes records after processing them, retaining transcripts for 90 days but allowing parental deletion requests.
This month, California approved a ban on manufacturing and selling toys equipped with conversational AI for children under 16 until 2031. The Toy Industry Association disputes this ban, advocating for comprehensive studies on risks to enhance safety guidelines.
Amidst these developments, “Miko 3” aims to penetrate Arabic markets with claims of supporting Arabic dialects. However, “Smart Life” found no independent safety assessments published concerning conversations in Arabic or details about the dialects tested, nor the targeted age groups. Parents require accessible information before purchasing such products, as children are unlikely to read usage terms while engaging with the toy.