abstract
- Meeting ABET's revised Outcome 4 has proven challenging, as paper-based case studies often fail to replicate the urgency and complexity of real-world engineering crises. This study employed a randomized controlled trial to evaluate a virtual-reality ethics module delivered through the VirtualSpeech¿ platform on Meta Quest 3 headsets, coupled with a large-language-model tutor providing adaptive formative feedback. Eighty fifthsemester mechatronic-engineering students were block-randomized-stratified by gender and grade average-into a VR-plus-AI intervention or a seminar-based control. Outcomes included a seven-dimension ABET Ethics Rubric, NASA-TLX workload ratings, and eye-contact analytics. Compared to the control, the VR-plus-AI group achieved statistically significant gains in ethical-decision quality (Hedges' g = 0.82), reduced perceived cognitive load, and enhanced gaze engagement, with all effects persisting four weeks post-intervention. These findings suggest that immersive realism combined with AI-driven feedback offers a scalable, evidence-based strategy for cultivating ethical competence and addressing accreditation demands in engineering education.