Artificial intelligence–enabled teaching: Insights from Kazakhstan higher education students

Abstract

The accelerated adoption of artificial intelligence (AI) in higher education has intensified expectations regarding instructional quality and learning effectiveness, yet empirical evidence on its pedagogical value remains contextually contingent. This study investigates students' perceptions of AI-enabled teaching in Kazakhstan, with a specific focus on perceived instructional quality dimensions and discipline-related skill development. Grounded in contemporary AI-supported learning literature, a structured self-administered questionnaire was developed and administered using a stratified random sampling design across higher education institutions. Instrument reliability and construct validity were established through Cronbach's alpha and the Rasch Rating Scale Model using a pilot sample representing 10% of respondents. Power analysis indicated a minimum requirement of 1504 observations; 2700 valid responses were ultimately analyzed. Data were processed using Microsoft Excel and RStudio, and partial least squares structural equation modeling (PLS-SEM) was employed to test seven hypothesized relationships within the proposed framework. The measurement model exhibited satisfactory reliability and validity, while the structural model showed low multicollinearity, strong explanatory and predictive power, and acceptable fit indices (SRMR, NFI, GoF). The findings reveal that usability, engagement, content quality, and accessibility of AI-based instruction significantly enhance student satisfaction, while instructional quality strongly predicts perceived skill acquisition, particularly in problem-solving, conceptual understanding, and technological competence. Conversely, perceived gains in interpersonal and diagnostic skills were comparatively weaker, and feedback-related pathways were not statistically significant, indicating limitations in current AI feedback mechanisms. The study offers robust perception-driven empirical evidence on the pedagogical implications of AI-integrated instruction in Kazakhstan's higher education system and provides actionable insights for evidence-based instructional design and AI-enabled teaching policy. • Power analysis established a minimum sample of 1504; the study analyzed 2700 valid student responses from Kazakhstan. • A large, demographically diverse dataset enhances the robustness and generalizability of findings on AI-enabled teaching. • PLS-SEM results demonstrate that usability, engagement, content quality, and accessibility significantly drive student satisfaction. • AI-supported instruction strongly predicts problem-solving, conceptual, and technological skill development, while feedback-related effects remain limited. • The findings provide actionable insights for higher education leaders and policymakers to optimize AI-driven teaching and digital strategies.

Ähnliche Arbeiten