Retroalimentación y ChatGPT en la educación superior. Revisión sistemática de literatura

Abstract

Currently, generative artificial intelligence has gained importance due to its use in feedback, particularly in universities. At the same time, it has sparked a debate about its influence on pedagogical dynamics. This systematic review seeks to provide insights into the discussion by evaluating the impact of ChatGPT in three key dimensions: students' academic training, teaching practices, and assessment. The PRISMA methodology was followed to select and analyze studies in Spanish (2019-2024) from SciELO, Redalyc, and Dialnet. The review included research evaluating automated feedback in higher education. The findings show that ChatGPT enhances content comprehension and promotes autonomy in structured disciplines but has limitations in fostering reflective analysis and logical argumentation. In teaching, it improves time management by reducing repetitive inquiries, although it still requires pedagogical mediation. In assessment, it is precise in identifying formal errors but lacks accuracy in conceptual judgments. In discussion the implementation of AI-driven educational approaches in Colombia raises concerns about technological disparities, the importance of continuous educator training alongside the absence of clear policy frameworks. To conclude is recommended to implement hybrid strategies, clear regulations, and AI training programs for teachers and students, ensuring equitable and efficient integration.

Ähnliche Arbeiten