Artificial intelligence in predoctoral research training: a proposal based on training practice

Abstract

The emergence of generative artificial intelligence in academic settings has sparked a deep debate about its implications for knowledge production, particularly during research training. This article presents an initiative developed within the Doctoral Program in Education at the University of Valencia (Spain), focused on the ethical, reflective, and formative use of tools such as ChatGPT, Copilot, Perplexity AI, and NotebookLM in the writing of doctoral dissertations. From a critical and situated perspective, the article analyzes the opportunities and dilemmas that generative AI poses for academic autonomy, authorship, and integrity. It also outlines curricular integration guidelines aligned with national and international policy frameworks. Through the description of a specific training seminar, pedagogical strategies are explored to train researchers capable of integrating these tools without compromising critical thinking or scientific integrity. This contribution aims to support the development of a responsible research culture that embraces critical thinking, advanced digital literacy, and knowledge ethics.