Abstract for: Leveraging K-12 Simulation Models and AI-based Coding to Create Fun Systems Thinking Games

Traditional education uses "static snapshots" that fail to teach nonlinear dynamics. Students struggle to learn from hard modeling tasks that become boring, while "black-box" edutainment hides underlying structures. Similarly, education systems struggle to integrate AI into learning. We apply a constructionist workflow (Kafai, 1995) using VS Code, Pygame, and Codex. Students use Codex to iteratively extract logic from expert-made SD models and author original games that effectively teach the original model's lessons, with some guidance.   This synthesis transforms students from passive analysts into system designers. By engineering games around rigorous systems thinking lessons, learners internalize complex feedback structures through an iterative process blending model testing and game debugging. We present Bank Dynamics Garage Edition: a game made in Codex and tested iteratively with Pygame inside VS Code in a few hours. Merging system dynamics with creative coding bridges the gap between abstract modeling and engaged learning. Using AI makes game creation easy and allows kids to focus on the pedagogical design of the model's intended lessons.