From corrector to mentor: the role of teachers in the era of generative AI

This systematic review addressed the integration of generative artificial intelligence (GAI) in higher education, a phenomenon that has redefined teaching, authorship, and assessment. Its purpose was to explore how the role of the teacher has transformed (from evaluator to mentor/designer), the benefits and risks reported, and the institutional and design conditions that have fostered authentic assessment. A pre-registered protocol was followed, including searches in Scopus, Web of Science, ERIC, SciELO, and RedALyC (2018–2025), followed by a three-stage screening process and an etic/emic content analysis based on pedagogical frameworks and levels of automation. The findings identified five teaching functions (feedback facilitator, engagement assessor, task designer, metacognition facilitator, and operational manager), benefits in writing, efficiency, and self-efficacy, and risks of dependency, loss of originality, and unreliable results without proper guidance. Clear policies, curriculum alignment, teacher training, and a process-oriented design (rubrics and traceability) emerged as critical factors. In conclusion, GAI has enhanced learning when used with pedagogical and ethical guidance; in its absence, it has amplified risks to academic integrity and quality.