AI-Supported Conditional Progression in Gamified Higher Education: Learning, Motivation and Cognitive Sustainability

Abstract

This paper presents an experimental study on gamified learning in higher education, comparing three learning conditions: traditional individual study, traditional gamification, and gamification based on AI-supported conditional progression. The study involved 142 university students, randomly assigned to the three conditions, who attended the same lecture, delivered by the course lecturer, and then worked on comparable learning content during a differentiated consolidation phase. The research adopts a cognitive ergonomics perspective, which looks beyond final performance and considers the learning experience as a whole, focusing on learning outcomes, motivation, perceived workload, and the cognitive sustainability of the path. In the control condition, students studied the materials individually, without gamified elements. In the traditional gamification condition, students moved through digital activities on a learning platform in a fluid way, without rigid constraints. In the conditional progression condition, students followed a structured digital path organised into levels, with immediate feedback, time limits, a maximum number of attempts, and advancement constraints linking each step to the previous one. Artificial intelligence functions supported the generation and organisation of the learning content, activities and feedback, always under human supervision: the researcher reviewed and validated the materials before their use in the classroom. No significant differences emerged among the three conditions in immediate learning (correct answers, p = .922; total score, p = .943) or overall perceived workload (p = .454), although the two gamified configurations produced different experiences. Traditional gamification showed higher perceived relevance (p < .05) and greater completion of required lessons than AI-supported conditional progression (M = 3.79 vs. 2.59; p < .001; d = 1.79). The latter offered more structure and guidance but shifted the focus of perceived demand toward managing progression rules, attempts and time. These results suggest that a more structured and technologically articulated design is not automatically a more effective or more sustainable one. The paper discusses implications for designing engaging, pedagogically coherent and cognitively sustainable AI-supported learning experiences, while future research should examine delayed learning and transfer and test less restrictive progression constraints across educational contexts.