Quantifying Student Success with Generative AI: A Monte Carlo Simulation Informed by Systematic Review

Abstract—The rapid rise of generative AI (GenAI) tools, such as ChatGPT, has intensified interest in their role in higher education and how students perceive and use them. This study combines a PRISMA-guided literature review with simulation-based modelling to examine student perceptions of GenAI. Nineteen empirical articles (2023–2025) were identified in Scopus; six reported item-level means and standard deviations suitable for probabilistic modelling. From this subset, one well-structured Likert dataset is selected as a canonical example to parameterise an inverse-variance-weighted Monte Carlo simulation, yielding a composite, perception-based “Success Score” (self-evaluated success). The simulation characterises both point estimates and uncertainty in this score under different thematic configurations. Within the simulated usability-derived themes, System Efficiency & Learning Burden receives the largest inverse-variance weight (highest precision) and therefore has the greatest influence on the composite perception-based Success Score under this weighting scheme. The approach provides a transparent and privacy-preserving bridge between thematic synthesis and probabilistic modelling for future work, linking GenAI perceptions to educational outcomes. Keywords—generative AI, Monte Carlo simulation, student success, usability, higher education