Play With AI (PL-AI): A Play-Centered, Design-Based Curriculum for AI Literacy in Pre-K and Kindergarten

Synthesis: This design-based research study reports on the design, implementation, and refinement of Play With AI (PL-AI), a developmentally appropriate curriculum introducing foundational AI literacy concepts to pre-kindergarten and kindergarten children. Two pre-K and two kindergarten teachers co-designed, piloted, and refined seven sequenced activities that progress from unplugged play (e.g., composing and testing "how-to" algorithms), to tangible coding (programming Bee-Bot or Ozobot), to guided dialogue with a social AI robot (Miko). Drawing on educational robotics, play-based learning, and the AI4K12 "Big Ideas" aligned with NAEYC's developmentally appropriate practice, PL-AI positions AI as a human-designed, instruction-following system rather than something autonomous or magical, embedding concepts like algorithms, data, and fairness within familiar centers, storytelling, and movement routines.

Key Findings

• Teacher co-design increased confidence and pedagogical Learner Agency — average self-ratings of comfort teaching AI topics rose across design cycles, with teachers attributing the shift to active participation in co-design and attributing curriculum ownership to collaborative activity-building (Teaching, Teacher Education).
• Four interrelated design principles emerged as feasible and sustainable: embodied play as conceptual grounding, tangible coding as a procedural bridge, guided dialogue as ethical and reflective Scaffolding, and teacher co-design as a Sustainability anchor.
• Children showed formative, multidimensional learning across conceptual, procedural, and dispositional strands — they began reasoning about AI as human-designed and instruction-following, developed sequencing and debugging skills through tangible coding, and displayed persistence, curiosity, and collaboration; ethical Reflection on fairness and human decision-making surfaced naturally through guided dialogue.
• Activities were refined through iteration — provisional prototypes like Data Detective and Sensor Safari were discontinued due to logistical constraints (their ideas absorbed into Scan and Draw and Color Code Stories), while spontaneous play like informal "races" was formalized into structured collaboration and debugging challenges.
• Implementation was embedded in existing classroom routines rather than treated as a stand-alone unit, with teachers mapping Bee-Bot routes onto number grids and alphabet mats and integrating AI exploration with literacy and numeracy goals.
• Findings are reported as formative design evidence, not validated outcome measures — the study was not designed to estimate effect sizes or establish causal learning gains, reflecting the epistemological commitments of design-based research.

What this means for practice

• Educators. Embed AI literacy inside existing centers, storytelling, movement, and literacy and numeracy routines instead of adding a stand-alone unit.
• Teacher educators. Use co-design with practicing teachers as the professional learning mechanism; participating teachers attributed their growth in confidence and curriculum ownership to building the activities themselves.
• Curriculum designers. Sequence activities from unplugged play to tangible coding to guided robot dialogue, and expect prototypes to be discontinued — Data Detective and Sensor Safari were dropped for logistical reasons.
• Administrators. Protect planning time and autonomy for teacher collaboration and iterative revision, which the study identifies as the Sustainability anchor for early childhood AI curriculum.

Limitations

• Design-based research with four teachers — two pre-kindergarten and two kindergarten — in two Title I schools in an urban district in the southern United States.
• The implementation lasted one month of PL-AI stations during freeplay (20–40 minutes), supported by 32 hours of classroom video, with no control group and no validated outcome measures.
• Evidence is teacher self-report, researcher field notes, observational rubrics, and participation tallies, which the authors present as formative design evidence rather than effect sizes or causal learning gains.
• Two of the seven activity prototypes were discontinued mid-study, so the activity sequences are context-dependent exemplars.

Citation

Lee, J. (2026). Play with AI (PL-AI): A play-centered, design-based curriculum for AI literacy in pre-K and kindergarten. Computers and Education: Artificial Intelligence, 10, 100569.