AI in Higher Education. Initial teacher training in the critical and didactic use of Artificial Intelligence

This study describes the application of a critical thinking model with artificial intelligence (AI) in the initial training of teachers at the University of La Laguna. The Educational Technology courses within the Bachelor’s degrees in Early Childhood Education and in Physical Activity and Sports Sciences (n = 103) were redesigned, integrating AI agents such as ChatGPT, Gemini and Perplexity into six phases for the critical use of these tools—questioning, comparison, critical dialogue, verification, re-elaboration and reflection—which guided students’ reflective interaction with AI. After the intervention, a mixed 23-item questionnaire was administered; descriptive statistics and thematic analysis of the “Evaluation and use of AI” dimension revealed that 88 % of students believe AI should be didactically promoted in higher education and 42 % identified ChatGPT as the agent providing the best answers. The most developed competencies were comparing outputs from different AI systems (66 %), designing effective prompts (63 %) and critically analysing responses (52 %). The main potentialities highlighted were rapid access to information and time saving, while perceived risks centred on plagiarism and cognitive dependence. The findings corroborate the pedagogical validity of the model for strengthening prompt-engineering skills and critical thinking, yet underscore the need to deepen ethical training and rigorous verification frameworks to address concerns about academic integrity and AI reliability. It is concluded that a reflective and regulated integration of these technologies can significantly enhance teacher education in higher education.