Harvard Study Examines How AI Toys Shape Child Development

Researchers at the Harvard Graduate School of Education are investigating how AI-enabled toys influence the cognitive and emotional growth of children. The study seeks to determine if these advanced devices serve as tools for development or create new risks for social behavior and attachment.

Led by Ying Xu, an Assistant Professor at the Harvard Graduate School of Education and affiliate of the Digital Wellness Lab, the project involves the creation of a custom AI toy. Xu and the team built the device specifically to begin testing with children in a laboratory setting. The research is supported by the Esther A. and Joseph Klingenstein Fund and is grounded in a conceptual framework detailed in the preprint "Artificial Intelligence in Toys: Implications for Child Development and Play," authored by Xu, Lauren Girouard, and Zhonghao Shi.

The Shift in Play

The study arrives as generative AI is increasingly integrated into physical consumer products. Historically, simpler AI toys—such as the early "Hello Barbie"—were criticized for using structured responses that could limit the range of stories children created. However, the emergence of generative AI allows for more fluid, unpredictable interactions. This shift prompts a critical question: does the AI act as a resource that extends a child's original ideas, or does it become an organizing force that narrows their imaginative choices?

Implications for Development

This research moves the conversation from theoretical debate to empirical evidence. By comparing interactions with AI toys against traditional non-AI toys, researchers can track how these devices influence language acquisition and social behavior. Of particular concern is the "redistribution of agency," or how the presence of an intelligent agent changes the way a child directs their own play. There are also concerns regarding whether these toys could foster unhealthy emotional dependencies or distort a child's understanding of truth and trust.

Future Research

Following the initial laboratory phase, the project aims to transition into longitudinal home-based studies. This next stage will allow researchers to observe how children interact with AI toys in natural environments over extended periods. By doing so, the team hopes to determine if AI toys eventually replace human interaction or if the cognitive benefits gained from these devices carry over into human-to-human social skills. The long-term impact on a child's sense of trust and the potential for AI-driven persuasion remain key areas for future observation.