Integrating AI-Based Image Analysis for Objective Assessment in CBET Practical Examinations

Abstract

Technical and Vocational Education and Training (TVET) in Kenya plays a vital role in equipping learners with practical skills essential for economic development. However, competency-based practical assessments in TVET institutions face challenges such as human bias, inconsistencies, and inefficiencies due to manual evaluation methods. This study investigated the potential of Artificial Intelligence (AI)-based image recognition to automate and enhance the accuracy, objectivity, and efficiency of CBET practical assessments. A mixed-methods approach was employed, including a baseline survey of 50 students and 10 trainers to assess AI awareness and perceptions, followed by the development and testing of an AI-powered image recognition system for evaluating AutoCAD practical tasks. Results indicate moderate AI awareness but high willingness for AI training among participants. The AI model demonstrated improved consistency and time efficiency compared to traditional instructor-led assessments. The findings demonstrate that AI-driven assessment tools offer fast and consistent evaluations, enabling timely feedback delivery even in large TVET classes. This scalability significantly reduces instructor workload while maintaining uniform standards across assessments. By facilitating fast, reliable, and scalable competency evaluations, AI can enhance skill development and employability in Kenya’s TVET sector, contributing to narrowing the digital divide through broader adoption of advanced technologies in TVET.