International Journal of Artificial Intelligence and Education Technology

Artificial intelligence (AI) is increasingly embedded in the administrative core of higher education institutions (HEIs), reshaping how academic leaders plan, allocate resources, and evaluate performance. Yet empirical evidence on whether AI-enabled decision-making actually improves institutional outcomes remains fragmented. This study examines the relationship between AI adoption, data-driven decision-making, perceived usefulness, trust and transparency, AI governance maturity, and institutional performance among administrators in Central Asian, South Asian, Southeast Asian, and MENA higher-education institutions. Using a cross-sectional quantitative design (N = 236), descriptive statistics, Pearson correlation, multiple linear regression, and one-way ANOVA were applied to test the hypothesized relationships. Data-driven decision-making emerged as the strongest predictor of institutional performance (β = 0.52, p < .001), followed by perceived usefulness (β = 0.29, p < .001) and AI governance maturity (β = 0.21, p < .001); trust and transparency contributed a smaller but significant effect (β = 0.13, p = .005). Notably, raw AI adoption lost statistical significance once these mediating constructs were controlled for (β = −0.05, p = .481), suggesting that the mere presence of AI tools does not translate into institutional performance gains unless it is channelled through disciplined, data-driven decision routines and sound governance. The full model explained 59.7% of the variance in institutional performance (R2 = .597, adjusted R2 = .588). These findings extend socio-technical models of educational technology adoption and offer HEI leaders an evidence base for prioritising governance and decision-culture investments over tool acquisition alone.

In the current era, the rapid evolution of artificial intelligence has opened doors to combining education with cutting-edge AI capabilities. However, prominent online learning platforms like Coursera and Udemy have not yet tapped into this synergy. This paper proposes an intelligent tutoring system that leverages AI innovations such as ChatGPT to enhance the online learning experience and improve student performance. To achieve this, the paper use the OpenAI interface to interact with ChatGPT, employing a suite of frameworks including React, Spring Boot, and Flask for a seamless front-end and back-end service delivery. On the infrastructure front, the robust capabilities of AWS, Kubernetes, Docker, and Jenkins is employed to facilitate continuous integration and continuous deployment (CICD). To assess the effectiveness of our system, we use a combination of A/B tests, tree tests, and questionnaires. These methods are complemented by stress tests and monitoring mechanisms to ensure reliability. While the proposed system satisfied the majority of functional and non-functional requirements, the experimental group demonstrated greater learning gains than the control group, with average scores increasing from 0.50 to 3.33 (improvement = 2.83), compared with 0.66 to 3.00 (improvement = 2.33) for the control group. The lower variance in the experimental group (0.333 vs. 1.00) indicates more consistent outcomes; however, the t-test results suggest that the difference in learning outcomes was not statistically conclusive.

Artificial intelligence (AI) technology is rapidly penetrating higher education, bringing profound transformations to university classroom instruction. However, existing research has paid limited attention to practical implementation pathways within real teaching contexts, leaving frontline educators without systematic operational guidance. Grounded in daily classroom practices, this paper aims to provide teachers with replicable and scalable strategies for Human-AI collaborative teaching, along with practical guidelines for applying AI tools. The study first outlines the contemporary context and driving forces behind integrating AI into university classrooms. It then systematically analyzes core challenges in current teaching practices—such as rigid teaching models, limited educational resources, inefficient classroom interaction, monotonous assessment methods, and excessive workloads for instructors. Finally, it proposes a comprehensive reform framework and implementation pathway for AI-enhanced university teaching, focusing on six key areas: reconfiguring teaching models, building intelligent resource-sharing systems, fostering student agency, establishing dynamic, whole-process evaluation mechanisms, supporting teachers in reducing workload and improving efficiency, and mitigating ethical risks associated with AI applications. Emphasizing practicality and feasibility, this paper offers frontline educators actionable solutions to navigate the transition toward intelligent education.