Generative artificial intelligence to measure critical thinking and academic performance in university students

Aspajo-Torres N.; Valdés-Cabodevilla R.C.; Hernández-Junco V.; Ziadet-Bermúdez E.I. · Revista de Ciencias Sociales (2026) · DOI: 10.31876/rcs.v32i3.46095

Resumen

Generative artificial intelligence has been incorporated into higher education, modifying academic production practices and demanding evidence regarding its relationship with critical thinking and performance. This study analyzes generative artificial intelligence as a tool for measuring critical thinking and academic performance in students at the Technical University of Ambato, Ecuador. A quantitative approach was used, employing applied research with a descriptive-analytical level and a non-experimental, cross-sectional design. Data collection was carried out using a five-point Likert scale questionnaire, comprised of 20 items distributed across four dimensions: critical reflection and evaluation; active search for evidence; epistemic awareness; and disposition toward independent thinking. The results showed high levels of critical thinking and significant, though limited, associations with academic performance. No systematic differences in performance were observed according to the tool used, except for a small effect size in the contrast between ChatGPT and Gemini. In the model, critical reflection and evaluation, active search for evidence, and epistemic awareness positively predicted performance, while disposition toward independent thinking was not significant. It is concluded that the contribution of generative artificial intelligence depends on guided use and verification practices, and longitudinal studies in the region are recommended. Licencia de Creative Commons Atribución 4.0 Internacional (CC BY 4.0) https://creativecommons.org/licenses/by/4.0/deed.es

Instituciones: Universidad Nacional Agraria de la Selva