The massive emergence of generative artificial intelligence tools —in particular large language models— has introduced a systemic variable that the post-Bologna pedagogical model could not foresee: the possibility that students obtain the products of academic work without having undergone the cognitive processes that such work was intended to foster. This article examines the implications of this phenomenon from three complementary perspectives. First, from cognitive neuroscience: neural plasticity operates on learning processes, not on their outcomes, so that the systematic cognitive offloading of tasks to AI impairs the development of the brain structures that sustain higher-order competencies. Second, from the available empirical evidence: the studies of Kosmyna et al. (2025), Gerlich (2025), and Yang et al. (2025) document significant negative effects of generative AI use on neural connectivity, critical thinking capacity, and higher-order reasoning in educational settings. Third, from the European normative framework on human oversight of AI: Article 14 of Regulation (EU) 2024/1689 makes effective human oversight a structural requirement of high-risk AI systems —a requirement that can only be met if professionals have previously acquired the substantive knowledge and reasoning capacity necessary to detect and correct system errors. The central paradox articulated in the article is as follows: a university education based on systematic cognitive outsourcing to AI produces supervisors who are incapable of supervising. In response, the article proposes a triad of operational methodological principles —sequence, redesign, and scale— illustrated with the author's own teaching experiences in administrative law and contrasted with emerging practices at North American universities. The proposal is not technophobic: AI is an amplifier of acquired competencies, not a substitute for their acquisition, and its integration into university teaching is desirable when it occurs at the appropriate stage of the learning process and under rigorous pedagogical criteria.
Eduardo Gamero-Casado (Thu,) studied this question.