Motivations, consequences, and concerns about the use of artificial intelligence in higher education: empirical evidence

Introduction The article analyzes the motivations, consequences, and concerns regarding the use of artificial intelligence (AI) among university students, with the aim of determining its influence on teaching-learning processes mediated in higher education. Methodology A quantitative analysis of the variables was conducted through the establishment of CHAID (Chi-square Automatic Interaction Detection) relationships among the motivations, consequences, and concerns perceived in students' educational dynamics. The analysis incorporated exogenous variables (academic workload, time pressure, sensitivity to rewards, and sensitivity to quality), endogenous variables (use of ChatGPT, procrastination, memory loss, and academic performance), and a set of sociodemographic factors (sex, age, employment status, academic cycle, marital status, previous educational sector, and parenthood). An intentional sampling was applied to 268 students from universities in El Salvador. Results The nine decision trees generated reveal heterogeneous yet consistent patterns. The most robust models show that academic cycle and perceived academic workload are the most relevant initial predictors, leading to pathways where difficulties in retaining information and intensive use of ChatGPT emerge as compensatory mechanisms. Discussion The findings open a problematic field concerning the diverse socioeducational processes and adjustments influencing higher education, confirming trends documented in previous international and regional studies. Conclusions The results highlight the urgent need to strengthen digital competencies, establish clear ethical guidelines, and promote a pedagogically oriented use of these technologies in Salvadoran higher education.