Co-design of Teacher Training in Artificial Intelligence and Evaluation in Uruguayan Secondary Education

Abstract

This paper presents the results of the co-design stage of the project “Strengthening Teachers’ Digital Competences,” focused on assessment mediated by Artificial Intelligence (AI) in 12 Centros María Espínola (CME) in Uruguay. The starting point was a previous mixed-methods diagnosis (N = 187 surveys and 11 focus groups) that revealed low competence levels in the area of assessment (M = 1.88 on a 4-point scale), along with marked ethical resistance to replacing the pedagogical relationship with algorithmic processes. In February 2026, collaborative workshops were held with the participation of more than 20 teachers and CME coordinators. The co-design methodology made it possible to identify that teachers demand situated training that does not focus solely on technique, but rather on the pedagogical redesign of tasks to foster critical paraphrasing and avoid uncritical plagiarism. Likewise, ethical red lines were defined, preserving human dimensions such as empathy and the socio-emotional context as domains exclusive to teachers. The results highlight the importance of Universal Design for Learning (UDL) and the seminar effect (peer learning) as pillars for the sustainability of training in contexts of vulnerability. This collective construction helps ensure that the final proposal, structured into flexible pathways, is feasible, responsive to real classroom needs, and coherently integrated into the Ceibal ecosystem.