Copyright (c) 2025 Revista Colón Ciencias, Tecnología y Negocios
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Artificial intelligence (AI) is transforming higher education, particularly in highly abstract disciplines such as mathematics. In this context, ChatGPT emerges as an innovative didactic resource capable of generating materials, guiding problem solving, and adjusting task complexity to students’ cognitive levels. This study conducted a scoping review following the PRISMA-ScR protocol and the PCC (Population, Concept, Context) framework, consulting twelve databases and institutional repositories for the 2012–2025 period. A total of 336 records were found. After applying inclusion and exclusion criteria (like language, relevance to the topic, and academic quality) and reviewing the studies using CASP, NOS, GRADE, and AMSTAR-2, twenty empirical and conceptual studies were selected. A narrative and categorical synthesis organized the findings into seven strategic axes: automated tutoring, content generation, AI-assisted instructional design, learning personalization, self-regulation, critical thinking, and teacher mediation. Reported benefits included enhanced autonomy and creativity, while the main challenges were conceptual inaccuracies, technological dependence, and ethical concerns. The PRISMA-ScR flow diagram summarized the selection process. Methodological limitations involved study heterogeneity and the restriction to English and Spanish publications. The review finds that using ChatGPT in university math teaching is important only if it is part of a critical teaching approach supported by school policies that promote fair access to technology, ethical behavior, and long-lasting improvements in education.
