ChatGPT and higher education in Latin America: measuring perceived academic skills

The use of generative artificial intelligence tools such as ChatGPT has transformed various academic practices in higher education, generating interest in understanding its perceived impact on the development of students' communicative and analytical skills. This study aimed to validate the Student Perception of ChatGPT instrument. A non-experimental quantitative design with an instrumental and comparative approach was applied to a sample of 2,579 students with prior experience using ChatGPT, from Ecuador, Mexico, Chile, Brazil, Colombia, and Guatemala. The questionnaire included 18 items organized into two key dimensions, communicative skills and analytical skills, measured on a five-point Likert scale. Internal consistency analyses (α = 0.943; ω = 0.944), as well as exploratory and confirmatory factor analyses, were conducted. The results confirmed high reliability and a coherent two-factor structure, with consistent groupings among items related to academic writing, digital communication, critical thinking, and problem-solving. These findings provide empirical evidence of the instrument's validity to measure students' perceptions of ChatGPT's contribution to the development of such skills. It is recommended to conduct longitudinal, mixed-method, and performance-based studies to more precisely examine its actual effect on university learning.