The Impact of AI-powered Character AI on EFL Learners’ Speaking: The Moderating Role of Anxiety

Document Type : Research Paper

Author

Allameh Mohaddes Nouri University

10.22099/tesl.2025.52912.3402

Abstract

Although several studies have examined the effectiveness of Artificial intelligence (AI) tools in enhancing L2 speaking, it remains uncertain whether learners with low and high levels of anxiety benefit similarly or not from AI-based instruction. The present study probed the degree to which learners with varying levels of anxiety, specifically high and low, benefit from an AI-based speaking assistant referred to as Character AI. This research involved 30 intermediate EFL learners from a language institute, who were identified and labeled as high anxiety (HA) and low-anxiety (LA) groups based on their anxiety scores. Following a sequential explanatory mixed-methods design, the initial phase involved administering an Oxford Placement Test (OPT) to determine the participants' language proficiency, followed by a pre-test that served as a benchmark for their speaking proficiency before exposure to the Character AI app. Over the course of the study, participants engaged with Character AI for speaking practice and feedback. The results disclosed that Character AI-based support proved beneficial in promoting FL speaking by fostering enhanced interaction and oral communication. It was also observed that learners with lower initial anxiety levels demonstrated greater improvements in speaking, implying that anxiety may moderate the benefits of AI-driven educational technologies. The qualitative findings underscore the potential of AI technologies as robust tools in L2 speaking, facilitating personalized, interactive, and learner-centric learning experiences. The present outcomes provide substantial insights into the interactions between psychological factors, technological integration, and language acquisition within the English as a Foreign Language (EFL) framework.

Keywords

Main Subjects



Articles in Press, Accepted Manuscript
Available Online from 08 July 2025
  • Receive Date: 23 April 2025
  • Revise Date: 29 June 2025
  • Accept Date: 08 July 2025