Efl Students’ Perceptions Of Ai Based Tools For Speaking Practice: A Mixed Methods Study

Razieq, Muhammad and Prihantoro, Prihantoro and Kusmaningrum, Melli (2026) Efl Students’ Perceptions Of Ai Based Tools For Speaking Practice: A Mixed Methods Study. Sarjana thesis, INSTITUT AGAMA ISLAM NEGERI CURUP.

Text
Muhammad Razieq FullText.pdf
Download (1MB)

Abstract

ABSTRACT Muhammad Razieq : EFL Students' Perceptions of AI based Tools for Speaking Practice: A Mixed Methods Study. Advisor : Dr. Prihantoro M.Pd Co-Advisor : Melli Kusmaningrum M.Pd The integration of Artificial Intelligence (AI) into English as a Foreign Language (EFL) instruction offers new opportunities for speaking practice, yet how students in regional Indonesian contexts perceive these tools remains underexplored; employing a sequential explanatory mixed‑methods design, this study involved 42 eighth‑semester students from the English Tadris Study Program at IAIN Curup who completed a questionnaire on their patterns of AI use, general perceptions, and perceived benefits and challenges, followed by semi‑structured interviews with six students selected through extreme case sampling to elaborate on the questionnaire findings. The results reveal that students utilize a diverse range of tools, most notably Duolingo speaking exercises (50.0%), ChatGPT voice/conversation mode (45.2%), and Google Translate voice input (40.5%), and generally hold positive perceptions (overall mean = 3.76), particularly valuing ease of use and expressing strong behavioral intention to continue using them, although perceived usefulness scored relatively lower; while students highlighted benefits such as flexibility, pronunciation support, and enhanced learner autonomy, significant challenges persisted, including speech recognition inaccuracies, limited access due to paid features, and the inability of AI to fully replace human interaction. The study further provides a nuanced tool‑specific mapping, revealing that students strategically combine different AI platforms for distinct purposes, which underscores the need for critical AI literacy and pedagogical guidance to maximize benefits while mitigating challenges, ultimately contributing empirical insights from an Indonesian regional university context and validating the Technology Acceptance Model in the domain of AI‑based speaking tools. Keywords: Artificial Intelligence, EFL Students, Perceptions, Speaking Practice, Mixed Methods Research

Item Type:

Thesis (Sarjana)

Creators: