Technology Acceptance of AI-Powered Student Registration Systems in Nigerian Polytechnics: A TAM/UTAUT Analysis

Manual student registration processes in Nigerian polytechnics regulated by the National Board for Technical Education (NBTE) remain a persistent administrative bottleneck, with each registration cycle consuming three to five working days per student and generating significant rates of data entry errors, document loss, and communication breakdowns across departments. This study assessed the technology acceptance of an AI powered Large Language Model (LLM)-based student registration system designed to automate document processing, form completion, and course validation tasks. Using an integrated TAM/UTAUT theoretical framework within a mixed-methods Design Science Research (DSR) approach, data were collected from 400 stakeholders (300 students and 100 staff members comprising lecturers and administrative personnel) at Federal Polytechnic Nyak, Shendam, Plateau State, Nigeria. Three validated pre-implementation questionnaires measured six constructs: Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Social Influence (SI), Facilitating Conditions (FC), Behavioural Intention (BI), and Trust (TR), using five-point Likert scales. Reliability analysis demonstrated strong internal consistency across all constructs, with Cronbach’s alpha values ranging from 0.783 to 1.000. Path analysis confirmed that all nine hypothesised relationships were statistically significant (p < 0.001): Perceived Ease of Use strongly predicted Perceived Usefulness (β = 0.752), while Perceived Usefulness was the dominant predictor of Behavioural Intention for students (β = 0.746). For staff, Facilitating Conditions emerged as the strongest predictor of Behavioural Intention (β = 0.602), reflecting the critical role of institutional infrastructure and technical support. Comparative analysis revealed that the only significant group difference was in Perceived Ease of Use (t (398) = 2.337, p = 0.020, d = 0.341), with students rating the system as easier to use than staff did. The findings provide strong empirical support for the deployment of LLM-based registration systems in Nigerian polytechnics and recommend differentiated training strategies that address the distinct acceptance drivers for students and staff. Implications for NBTE accreditation policy, institutional digital transformation planning, and future research on AI adoption in African higher education administration are discussed.