A Systematically Informed Conceptual Review of the Effect That Generative AI Has on University Students’ Reported Self-Efficacy

Abstract

The consistent growth in the use of Generative artificial intelligence (AI) globally has led to an increased focus on attempts to measure the impact that it has on university students, including their academic motivation and ability. An emerging interest in the literature addresses the impact that the use of Generative AI may have on students’ reported self-efficacy—and specifically whether use of Generative AI is associated with an increase or decrease in reported self-efficacy. This systematically informed conceptual review draws on the theoretical construct of self-efficacy in analyzing papers published in 2025 and 2026. The results suggest an initial framework that identifies the factors that determine whether university students’ use of Generative AI is likely to enhance or diminish their self-efficacy. The analysis presented here identifies the following factors as important determinants of the impact that Generative AI usage is likely to have on students’ self-efficacy: pedagogic support (the extent to which the Generative AI is introduced with learning and teaching support); task usefulness (whether the Generative AI has the quality and relevance to address the specific focal task; and unique affordances (whether the Generative AI has qualities that make it uniquely valuable in addressing the task). Consistent with the Triple Helix framework, implications for governments, industry, and educators regarding the importance of self-efficacy in education policy, in AI resource development, and also in the pedagogic framework of its deployment with students are identified. Furthermore, the relevance of these findings for Sustainable Development Goal 4 (SDG 4) is also identified.