Generative Artificial Intelligence (GenAI) is shaping higher education, yet its adoption does not necessarily demonstrate improvements in teaching and learning. This matters in Civil Engineering, where innovation must preserve technical reliability, professional judgement, ethical responsibility and human oversight. This paper proposes a performance and assurance framework to help Civil Engineering educators evaluate GenAI-supported teaching practices. Starting with an identified teaching need, it maps existing and AI-supported workflows, identifies stakeholders and measures pedagogical value, feedback quality, consistency, staff efficiency and usefulness. Performance is assessed alongside assurance requirements addressing technical accuracy, hallucination risk, human validation, accountability, transparency, data governance, ethical use and sustainability. The framework’s distinguishing feature is its decision logic: critical assurance requirements operate as non-compensatory constraints, so efficiency gains cannot justify technically incorrect or educationally unsuitable outputs. Applicable across Civil Engineering teaching activities, it can be illustrated through GenAI-assisted feedback on hydrostatic force calculations. Feedback can be assessed for pedagogical clarity, identification of conceptual errors and lecturer preparation time, while technical accuracy is checked against an expert-verified solution. Failure to meet essential safeguards requires modification or rejection, regardless of performance gains. Embedded within a Plan–Do–Check–Act cycle, this evaluation supports explicit decisions to adopt, modify or reject GenAI-supported practices and revisit them as needs evolve. The framework offers an educator-led, discipline-specific approach to determining whether GenAI delivers meaningful educational improvement rather than simply increasing its use. Decisions remain transparent, repeatable and professionally grounded.
Antonia Chatzirodou is a Senior Lecturer and Programme Leader in Civil Engineering at the University of the West of England (UWE Bristol). She holds a PhD in Coastal Engineering and specialises in numerical modelling, water and environmental engineering, and engineering education. Her interests include digital innovation, responsible Generative AI, and evidence-based approaches to improving teaching and learning. She is committed to advancing engineering education while maintaining technical reliability, professional judgement and ethical responsibility.
Professor Giacinto Barresi
Dr Marianthi Leon