LINGUISTIC AND VISUAL PATTERNS OF CHATGPT IN HIGHER EDUCATION An Analysis of its Use in Undergraduate and Postgraduate Theses

Peña-Cáceres O.; Sánchez-Rogel E.; Barros-Naranjo J.; Silva-Marchan H.; Espinoza-Nima R.; Correa-Calle T. · VISUAL Review. International Visual Culture Review / Revista Internacional de Cultura (2025) · DOI: 10.62161/revvisual.v17.5403

Resumen

This article examines the influence of artificial intelligence on academic writing, focussing on undergraduate and postgraduate theses. Its objective is to identify linguistic and visual patterns in texts generated or assisted by ChatGPT through an analysis of 50 theses, equally distributed between a public and a private university in Peru. The results indicate that the most representative argumentative structures include terms such as “moreover,” “crucial role,” and “comprehensive.” A 48% argumentative similarity was observed in the public university, in contrast to 72% in the private university. It is recommended to broaden the analysis by incorporating other language models, such as BART.

Instituciones: Universidad Nacional de Piura; Universidad Tecnológica del Perú; Universidad Nacional de Tumbes