An Explainable Hybrid AI–Gamification Framework for Adaptive E-Learning Systems with Simulation-Based Validation

Gamified e-learning platforms and Artificial Intelligence (AI)-driven adaptive learning systems can improve learner engagement, but existing solutions rarely combine the two, and few explain their adaptive decisions to learners. To address this gap, we propose LearnSpace, a hybrid AI–gamification framework for responsive e-learning that adds an Explanation Layer, based on Explainable Artificial Intelligence (XAI), to a deterministic reward architecture. The framework merges classical gamification mechanics with an AI personalization module in which experience points, rewards, and difficulty adjustments follow explicit weighted formulas. A Thompson-sampling bandit learns which feedback works best after a mistake. For each adaptive decision, the Explanation Layer names only the factors that drive it and turns them into a short justification for the learner. The design draws on a qualitative overview of 14 platforms, and the LearnSpace mobile prototype implements the gamification components and the bandit. In a pre-registered simulation with 500 synthetic learners per run, the first explanation rule named a factor that had no effect on the decision in 95.3% of difficulty adjustments, while the decision-specific rule never did. The bandit had lower regret than uniformly random feedback allocation. Effects on real learners remain to be tested.