A Replicable Pipeline and Framework for Developing Educational Serious Games Using a Generative Artificial Intelligence Platform

• Grant number: 26/18213-0
• Support Opportunities: Research Grants - Regular Innovation Grant - AIR
• Start date: October 01, 2026
• End date: March 31, 2028
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Eloíza Martins Primo Capeloci
• Grantee: Eloíza Martins Primo Capeloci
• Host Institution: FAC TECNOLOGIA POMPEIA SHUNJI NISHIMURA/CEETPS

Abstract

This project proposes the development, validation, and systematization of an artificial intelligence-assisted authoring pipeline for educational serious games, implemented through a generative game-development platform. The research will be conducted at Fatec Pompeia Shunji Nishimura and will use courses from the Precision Agriculture Mechanization Technology and Intelligent Systems Technology programs as its application and validation settings.The project addresses the technical barriers that prevent teachers without programming expertise from developing educational games with meaningful mechanics, such as simulation, problem-solving, and decision-making. The production of these games usually requires specialized development teams, making the process costly and time-consuming. In contrast, accessible no-code tools are generally limited to quizzes and do not adequately support the procedural and applied skills required in technological education. Although generative artificial intelligence can transform syllabi, presentations, and natural-language descriptions into functional games, systematic methods are still needed to ensure pedagogical quality, mechanical diversity, technical reliability, and safe content generation.The research will adopt a Design-Based Research approach, organized into iterative cycles of design, implementation, analysis, and refinement. The project will develop prompt-engineering procedures, generation contracts, templates for different game mechanics, objective quality criteria, and guardrails to reduce errors and artificial intelligence hallucinations. The performance of the proposed pipeline will be compared with manual game-development processes based on authoring time, rework, cost, variety of mechanics, error rates, and the proportion of outputs that can be used without manual code correction.The resulting method will be consolidated into a replicable framework comprising a playbook, templates, validation protocols, and an adoption kit that can be transferred to other courses, Fatecs, and Etecs. The framework will be empirically evaluated with students and teachers, considering usability, engagement, adoption effort, and learning indicators. Expected impacts include reducing the time and cost required to produce serious games, expanding teachers' access to active learning resources, and delivering an open, measurable, documented, and scalable method for the responsible use of generative artificial intelligence in technological education. (AU)