Wonder Town: Integrating Generative AI into a Bloom-Aligned Physics Learning Game

Key insights

6 takeaways

01
Wonder Town introduces a multi-touchpoint design pattern that integrates generative AI throughout a seven-phase learning loop based on Bloom's Taxonomy, rather than using AI as an isolated chatbot or narrative engine.See it on page 2
02
The game structure maps every chapter to Bloom's levels—Remember, Understand, Apply, Analyse, Evaluate, and Create—to ensure cognitive progression from foundational knowledge to higher-order thinking.See it on page 2
03
To maintain pedagogical integrity, the system keeps physics simulations and quiz content hand-coded, using generative AI only for narrative flavor and guided reflection to prevent hallucination and curriculum misalignment.See it on page 5
04
Formative evaluation with students and teachers revealed that direct-instruction formats were perceived as 'boring,' leading to a redesign that embeds physics concepts within personalized, narrative-driven town environments.See it on page 3
05
Preliminary mixed-methods results from five students (ages 12-15) showed that all participants reported increased interest in physics, with 60% successfully applying learned concepts to real-world scenarios outside the game environment.See it on page 4
06
The system allows learners to choose a narrative style (e.g., Sci-Fi) at the start of a session, which triggers a one-time generation pass that updates the story prose while keeping the underlying physics and learning objectives fixed.See it on page 3

Summary

Wonder Town: Integrating Generative AI into a Bloom-Aligned Liangjun Fu Aurora Constantin School of Informatics School of Informatics University of Edinburgh The University of Edinburgh Edinburgh, United Kingdom Edinburgh, United Kingdom [email protected] [email protected] Gennaro Imperatore Kejia Zhang Computer and Information Sciences Edinburgh Parallel Computing Centre University...

Wonder Town: Integrating Generative AI into a Bloom-Aligned Physics Learning Game

Wonder Town: Integrating Generative AI into a Bloom-Aligned Physics Learning Game

Key insights

Summary

Next steps

More from University of Edinburgh

Human-Agent Trust Relationships in a Real-Time Collaborative Game

Games-Based Learning as an Interdisciplinary Approach to Literacy across Curriculum for Excellence

More on AI

Fostering Computational Thinking in Children with ASD: A Tailored Digital Game Design Proposal

Extending The Project Win Game: Teaching Real Options and Moral AI Project Decision-making

Mission Profiles and Interaction Patterns in Gravitas: Evidence from a Simulation-Based Physics Learning Game

An Early-Stage Feasibility and Usability Pilot for an Immersive Serious Game to Support English Spelling and Semantic Skills

Learning AI Ethics through Game Design as Ethical Inquiry

Emoji Story Prompts: LLPx 2025 Symposium Post-Conference Publication

A Serious Game to Study Reduced Field of View in Keyhole Surgery: Development and Experimental Study

Design of Cryptopol: A Serious Game for Teaching Cryptocurrency Tracing Techniques to Law Enforcement

Framework-based Roguelike Game for AI/ML Education

Bayesian Knowledge Tracing Implemented in a Telecommunications Serious Game

LingoTowns: A Virtual World for Natural Language Annotation and Language Learning

The Crypt of Notation: Rote Learning through Video Games for Adult Beginner Keyboard Learners