Abstract

This study examines an artificial intelligence (AI)-supported mathematical reasoning model for biological systems modeling and engineering education. In particular, it examines the effects of a PBL framework on undergraduate engineering students solving modeling problems that also include the use of an algorithm designed to aid mathematical reasoning. The study was conducted using a quasi-experimental design featuring a control group (n = 35) and an experimental group (n = 36) drawn from 71 students in the Mathematical Foundations course at Universidad Técnica Particular de Loja. Assenting to group equivalence at baseline (Levene's test; Student's t-test: t = 0.425; p = 0.672), both groups participated in an academic term using an AI-mediated PBL approach. Positive post-intervention test results reveal a statistical advantage for the experimental trials (M = 8.5) over the control trials (M = 7.4) in statement interpretation (t = -2.295; df = 69; p = 0.012) and in algebraic modeling (t = -2.360; df = 69; p = 0.015). Graphical representation also improved (p = 0.042), but it remains the most challenging area. The results highlight the value of AI as a ‘cognitive scaffold’ that does not replace teacher mediation but complements it to enhance learning, and they have implications for how curricular content is governed to foster digital literacy and ensure proper use of AI for learning in the engineering domain.