79.3% of prospective teachers who participated in an educational experience involving ChatGPT checked its responses against other sources and modified them before using them. This is one of the key findings of a study by Ana María Pinto-Llorente, Vanessa Izquierdo-Álvarez, and Marta M. Dolcet-Negre, published in RIED, which examines how 174 students in teacher education programmes at the University of Salamanca perceive generative artificial intelligence.
Following three practical sessions devoted to designing learning situations for early childhood and primary education, participants rated the tool very positively: satisfaction reached 4.45 out of 5, while perceived usefulness scored 4.24. Yet only 15.5% used the generated responses directly without making any changes.
The study reveals a combination of technological enthusiasm and critical awareness. The highest scores were recorded for concern about the ethical and social impact of AI (4.34), perceived barriers to its integration into teaching (4.18), and recognition of its educational relevance (4.08). Using correlation network analysis, the researchers identified three clusters of dimensions (knowledge, evaluation, and obstacles) with confidence and perceived relevance emerging as two of the most highly connected factors.
Those who feel more confident using these tools and recognise their pedagogical value also tend to report stronger practical skills and higher expectations of future use. Master's students, moreover, show greater awareness of the ethical and social implications of AI than undergraduate students.
One of the clearest implications concerns initial teacher education: 67.2% of participants believe that both teachers and students need specific training in generative AI, while a further 27.6% consider such training necessary for teachers alone. The authors recommend combining hands-on experience, critical evaluation of AI-generated content, ethical reflection, and guided support to strengthen teachers' professional confidence.
The research, based on self-reported perceptions from a sample drawn from a single university, cannot establish causal relationships or demonstrate objective improvements in the quality of learning situations. It does, however, show that a positive attitude towards AI can coexist with critical scrutiny of its outputs and an awareness of its limitations.
---
How to Cite: Pinto-Llorente, A. M., Izquierdo-Álvarez, V., & Dolcet-Negre, M. M. (2026). Self-perception and usefulness of generative artificial intelligence among pre-service teachers. RIED-Revista Iberoamericana de Educación a Distancia, 29(1), 111–132. https://doi.org/10.5944/ried.45480
