Deep learning in vocational simulations: teacher transformation in third-year general unified baccalaureate

Abstract

This study analyzes the impact of a deep learning–based vocational simulation on the transformation of vocational guidance practices among teachers in the third year of General Unified Baccalaureate in Cuenca, Ecuador. A mixed-methods, quasi-experimental, one-group pretest–posttest design was implemented, combining open educational data from the Ministry of Education and “Cuenca en Datos” with teacher and student questionnaires, simulation usage logs, and qualitative techniques (interviews and focus groups). The sample comprised 40 teachers and 960 students from 11 schools (public, municipal, and private) in both General and Technical Baccalaureate tracks, within the broader implementation of the National Digital Education Agenda and AI-driven digital transformation strategies in the Ecuadorian school system. Results show significant increases in the use of academic data and digital resources in vocational guidance, as well as in teachers’ perceived competence to use AI tools in a pedagogical and ethical manner. From the students’ perspective, the intervention led to higher clarity regarding life projects, reduced vocational uncertainty, and a marked rise in interest in digital and STEM careers, in line with recent evidence on AI in vocational guidance programs. The study concludes that integrating deep learning–based vocational simulations with open data, supported by teacher professional development in AI, can catalyze the transformation of vocational guidance in upper secondary education, aligning it with national graduate-tracking policies and international recommendations on the responsible use of AI in education.