his paper examines curriculum viability in Civil Engineering as an architectural problem rather than a purely disciplinary one. Using disruptive search campaigns within the AlphaPath framework, it compares alternative curricular architectures and shows that content modernisation alone does not improve viability if subject-fragmented serial logic remains intact. The winning design combines integrated studios, distributed closure, and human-verified AI-supported modelling, achieving substantially higher graduation performance than both the 2005 legacy curriculum and its best repair-based variants. The paper also proposes a taxonomy of non-delegable competences for Civil Engineering education in the age of artificial intelligence.
Hugo Roger Paz (Mon,) studied this question.