Environmental transformation in the classroom: evaluation of a narrative GPT effect on university students’ environmental awareness

Purpose This study aimed to evaluate the impact of a personalized GPT, Alma de Tezontle, on the environmental awareness of university students. The GPT was designed using a validated personalization methodology grounded in life stories and dialogic interaction to foster empathy and recognition. The objective was to assess whether this narrative AI model could generate measurable changes in students’ beliefs, behaviors and concern regarding the environmental crisis, particularly water pollution and to identify gender differences in its effect. Design/methodology/approach A quasi-experimental design was implemented with 92 students from a university in western Mexico. Participants were divided into a control group (n = 32) and an experimental group (n = 60), which interacted with the GPT. The validated instrument adapted from the Andalusian Ecobarometer measured seven dimensions of environmental awareness. Data were analyzed using permutation-based nonparametric tests (maxT statistic) in R to evaluate significant changes by group and gender. Findings Statistically significant changes were found in three dimensions: reduction of minimization beliefs (p < 0.0001), increase in pro-environmental behaviors (p = 0.0003) and heightened environmental concern (p = 0.0031). Gender differences emerged: men showed more cognitive shifts (belief reduction), while women expressed more behavioral changes and concern. Research limitations/implications The sample was limited to one institution and a short-term intervention, restricting generalizability and long-term inference. The analysis relied solely on quantitative data, limiting insight into participants' subjective experiences and the mechanisms behind attitudinal change. Originality/value This is one of the first empirical studies to evaluate a personalized GPT as a transformative educational agent. Its originality lies in combining narrative AI with environmental education, using a validated methodology to generate emotional engagement and measurable behavioral change.