Predictive AI for Academic Performance: Integrating Natural Language Processing, Gamification and Self-regulated Learning

Artificial Intelligence and Natural Language Processing are widely used to predict academic performance across education levels. This study examines how students’ personality and maturity relate to academic progress by comparing their actual degree program year with predictions generated through advanced Artificial Intelligence and Natural Language Processing techniques. Our predictive model achieved a mean absolute error of 0.794. Maturity and personality traits were derived from two sources: (a) Type-Token Ratio and Flesch-Kincaid Grade used to assess developmental writing features; and (b) Hexad Model player-types to classify motivational traits. Both data types were collected through post-task surveys following a gamified activity. The model was applied to students across all four years of a Tourism degree program. Results show that NLP-based linguistic features and gamification profiles were the strongest predictors. The data confirmed expected gains in lexical diversity and readability across academic years. Additionally, shifts in player types – from first to fourth year – suggest evolving motivational orientations and personality development. These findings offer valuable insights for identifying students whose developmental trajectory may not align with their academic standing. When predicted and actual program year differ, educators can use this signal to provide targeted, timely support. AI-based writing analysis fosters maturity monitoring, metacognitive awareness and self-regulated learning. Meanwhile, personality profiling enables differentiated instruction based on motivational drivers. This methodology offers a scalable tool for inclusive, personalized education – particularly in multilingual settings where accurate translation preserves students’ linguistic voice.