GenAI Playground
A Study of Generative AI in
Co-Creative Constructionist Learning Environments
Youth Development Program -
Research Project
Research Fellows
Dr. Zahra Hejrati, Dr. Hanieh Khaleghian
Project Summary
The rapid emergence of generative AI presents a critical choice for the future of childhood learning: will these tools function as “instructionist” tutors that simply deliver answers, or as “constructionist” partners that engage children in meaningful creation? Drawing on Seymour Papert’s and Mitchel Resnick’s foundational work – notably the idea that children enjoy “hard fun” and willingly tackle challenging tasks when they care about the project – this research proposes to design and study Gen AI Playground, a flexible digital environment for children (ages 7–12) to create with AI. The Playground will host a suite of creative AI partners (e.g., a Story Spinner or Game Builder) designed not as answer-givers, but as collaborators that support a child’s agency and engagement in open-ended projects. Instead of doing tasks for the child, these AI partners will enable them to develop their own ideas. Through an observational study in this environment, we will investigate how different AI interaction paradigms affect children’s creative process, persistence, and sense of authorship. The primary goal is to develop empirically grounded design principles for AI tools that amplify, rather than automate, human creativity in learning.
Expected Outputs Duration
Concretely, this research will yield several valuable outputs:
● Design Principles for Constructionist AI: A set of empirically grounded design guidelines for creating AI tools that act as creative learning partners. For example, principles might include how an AI should phrase its prompts or reveal its process to maximize the child’s creativity and sense of ownership. We will distill lessons on balancing helpful scaffolding with open-ended exploration, providing a reference for developers of future educational AI software.
● Research Report: Rich qualitative and quantitative findings on how children perceive and build a relationship with an AI partner. By documenting children’s behaviors and attitudes in our study, we contribute a much-needed child’s perspective to debates on AI in education. This includes understanding the language children use to describe AI’s role, the conditions under which they either flourish or become frustrated, and how their creative trajectories unfold with varying levels of AI support. Such insights will be shared through publications and workshops, informing both academic discourse (e.g., in HCI and learning sciences) and practical training for teachers on the frontline of integrating AI tools.
Duration:
January 2026 – December 2026, 12 months
Research Advisors:
Dr. Yahya Tabesh, Dr. Mohsen Hejrati