UPDF AI

Generative Artificial Intelligence in Education: Learning Experiences through Game Development

Marcos Monte de Oliveira Alves,Priscila Marques Dias Corrêa,Filipe Barcellos dos Santos

2026 · DOI: 10.69906/gepem.2176-2988.2025.1276
Boletim GEPEM · 0 Citations

TLDR

This study proposes using ChatGPT as a leading agent in coding a 3D 4x4x4 Tic-Tac-Toe game, and employs a methodology to interact with it, while analyzing its autonomy in light of the challenges faced and students’ learning outcomes throughout the process.

Abstract

The use of generative artificial intelligence has been debated for the past few years. Motivated by game development, this study proposes using ChatGPT as a leading agent in coding a 3D 4x4x4 Tic-Tac-Toe. The main goals of the study were to comprehend whether ChatGPT could develop the game code with minimal aid and to what extent students could enhance their computational thinking skills when engaged in this process. Inspired by Papert’s work, this study presents an appealing collaboration. It combines the computational power of generative artificial intelligence with human intervention, providing a rich and productive learning environment. The study employs a methodology to interact with ChatGPT, while analyzing its autonomy in light of the challenges faced and students’ learning outcomes throughout the process. Findings suggest that ChatGPT has the potential to develop games mostly independently. Yet, human intervention is still needed, and knowledge-building is still evident.