AI education researcher Randi Williams said parents should encourage children to question and even dismantle AI-embedded toys to foster critical thinking.
This approach addresses the growing emotional bonds children form with artificial intelligence. As these devices become more integrated into early childhood development, the risk of unquestioned reliance on algorithmic companionship increases.
Williams discussed these concerns during a TED Talk at the Play@TED conference event on May 14, 2026 [1]. She said that AI-enabled toys are designed to create strong emotional ties with their users. While these bonds can feel genuine to a child, they are products of programmed responses rather than human empathy.
To counter this, Williams proposes that children engage with the technology from a position of curiosity and skepticism. By teaching kids to ask how a toy works or why it says a specific phrase, parents can shift the child's role from a passive consumer to an active investigator.
Williams said that the act of physically or conceptually dismantling the technology is a powerful educational tool. This process helps children understand that the "personality" of the toy is a result of engineering, not a sentient being. Such transparency allows children to maintain a boundary between human relationships and machine interactions.
The researcher said that this critical engagement prevents the blind acceptance of AI outputs. When children learn that technology can be questioned or broken, they develop the agency needed to navigate a world saturated with automated systems. This method transforms a potential psychological dependency into a lesson on digital literacy.
“AI-enabled toys are designed to create strong emotional ties with their users.”
The shift toward 'adversarial' play—where children are encouraged to challenge or break technology—represents a move away from traditional educational models that emphasize seamless user experiences. By prioritizing critical literacy over emotional attachment, researchers aim to protect the cognitive development of children against the persuasive nature of generative AI.



