Play with AI (PL-AI): A play-centered, design-based curriculum for AI literacy in pre-K and kindergarten

Key findings from this study

• The study found that teacher co-design participation substantially increased pedagogical confidence, curriculum ownership, and reflective practice compared to pre-implementation baseline conditions.
• The researchers demonstrate that four interrelated design principles—embodied play, tangible coding, guided dialogue, and teacher co-design—emerged as both feasible and sustainable for embedding AI literacy in early childhood classrooms.
• The authors report that children engaged in meaningful reasoning about algorithmic and computational concepts through unplugged activities, robot programming, and comparative dialogue with a social AI robot.

Overview

Play with AI (PL-AI) is a developmentally appropriate curriculum introducing foundational artificial intelligence concepts to pre-K and kindergarten children through play-centered design. The curriculum integrates unplugged algorithmic play, tangible coding with programmable robots, and guided dialogue with a social AI robot. Four teachers co-designed, piloted, and iteratively refined seven sequenced activities across multiple implementation cycles.

Methods and approach

The study employed design-based research methodology with four early childhood teachers (two pre-K, two kindergarten). Data collection included teacher surveys, 32 hours of classroom video recordings, researcher field notes, and design-meeting transcripts. Teachers participated in collaborative design and refinement cycles, with analysis focusing on enactment patterns, teacher adaptation, and children's emerging reasoning about AI concepts.

Results

Teachers' confidence and pedagogical agency increased substantially through co-design participation, with ownership of the curriculum and reflective practice deepening across implementation cycles. Four interrelated design principles emerged as both feasible and sustainable: embodied play, tangible coding, guided dialogue, and teacher co-design. These principles supported conceptual clarity, developmental appropriateness, and long-term sustainability of AI literacy instruction in early childhood settings.

Classroom enactment demonstrated meaningful teacher adaptation to local contexts and student needs. Children engaged in reasoning about algorithmic processes through unplugged activities, programming tasks with Bee-Bot and Ozobot robots, and comparative dialogue with the social AI robot. Iterative refinement cycles produced increasingly aligned implementation of curriculum activities with intended pedagogical aims.

Implications

The study contributes empirically grounded design knowledge for early AI education aligned with AI4K12 Initiative and NAEYC frameworks. Results suggest that teacher co-design processes substantially enhance implementation feasibility and pedagogical sustainability, warranting integration of collaborative design into professional development for AI literacy initiatives. The four emergent design principles provide practical guidance for educators integrating AI concepts through play-based, ethical, and collaborative approaches in early childhood classrooms.

Findings indicate that AI literacy need not depend on computational tools alone; unplugged play and tangible coding with accessible robots support conceptual understanding in developmentally appropriate ways. The emphasis on guided dialogue and teacher agency suggests that successful early AI education requires ongoing reflection and adaptation rather than standardized scripted implementation. Future research should examine how these design principles scale across diverse early childhood contexts and communities.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.