Abstract:

The accelerating influence of the Fourth Industrial Revolution (4IR) has transformed traditional classrooms into data-rich, sensor-enabled smart learning environments. With Internet of Things (IoT) technologies generating continuous streams of behavioural and interaction data, and Artificial Intelligence (AI) offering advanced analytical capabilities, new opportunities exist for enhancing the validity and fairness of educational assessments. Yet, many assessment systems still rely on static, uniform test formats that overlook the depth of information captured in technology-enhanced classrooms. This study proposes an integrated IoT–AI psychometric framework designed to improve learning outcome assessments in alignment with Sustainable Development Goal 4 (SDG 4). Data were collected from 312 undergraduates across IoT-enabled classrooms, combining sensor logs, LMS activity, and computer-based test performance. AI models, including artificial neural networks and random forests, were utilised to classify engagement and predict learning outcomes, while Item Response Theory (IRT) models were used to calibrate assessment items. Findings indicate that IoT indicators, such as time-on-task and interaction frequency, strongly predict academic performance. AI models achieved high accuracy (up to 92%), and integrating behavioural features into psychometric models improved item discrimination and reduced Differential Item Functioning (DIF) by 17–26%. The results demonstrate the potential of smart assessments to deliver more accurate, equitable, and context-aware evaluations, advancing progress toward inclusive, high-quality education under SDG 4. The findings demonstrate that combining IoT behavioural data with AI-powered psychometrics produces more accurate, equitable, and context-aware assessments than traditional methods. This integrated approach strengthens diagnostic feedback, improves measurement validity, and supports SDG 4’s vision of inclusive and high-quality education. The study provides a scalable foundation for next-generation smart assessments in developing and developed contexts.

Keywords:

Smart classroom, smart assessment, IOT, psychometric evaluation, SDG 4.