The Impact of Generative AI on Self-Efficacy: Evidence From a Meta-Analysis
TL;DR
The findings suggest that genAI can support learners’ self-efficacy when it is used to scaffold learning and encourage active engagement, and highlight the importance of how genAI is integrated into instruction when designing AI-supported learning environments.
Abstract
Generative artificial intelligence (genAI) tools such as ChatGPT are increasingly used in education. While many studies have examined how genAI tools affect learning, we know less about how genAI tools influence motivational factors such as learners’ self-efficacy, despite the central role self-efficacy plays in motivation and learning. We conducted a meta-analysis of experimental and quasi-experimental research to examine the effect of genAI tools on learners’ self-efficacy. Studies published between November 2022 and February 2026 were systematically searched and screened, yielding 84 independent effects in our analysis. The results showed that using genAI during learning improved learners’ self-efficacy with a moderate effect size. However, the size of this effect depended on how genAI was used. Implementations where genAI acted as a tutor that guided students’ learning produced larger improvements in self-efficacy than uses where AI only (1) provided feedback, (2) answered questions, or (3) replaced learner work. Also, longer interventions produced stronger improvements than brief interactions. The improvements in self-efficacy did not differ across academic domains or learner age groups. These findings suggest that genAI can support learners’ self-efficacy when it is used to scaffold learning and encourage active engagement. The results highlight the importance of how genAI is integrated into instruction when designing AI-supported learning environments.