EFL SAUDI UNIVERSITY LEARNERS' PERCEPTIONS AND USE OF GENERATIVE AI SOFTWARE IN LANGUAGE LEARNING

Abstract

As generative artificial intelligence (AI) tools become increasingly integrated into educational ecosystems, understanding how learners interact with these technologies is crucial, particularly in underrepresented English as a Foreign Language (EFL) contexts. This dissertation investigates the perceptions and usage patterns of Saudi university students as they interact with generative AI software to support their English language learning. While AI technologies like ChatGPT, Grammarly, and QuillBot are rapidly transforming academic practices, limited research has explored their impact from the learner’s perspective in non-Western, non-native English-speaking environments. This mixed-methods study, grounded in the Technology Acceptance Model (TAM), was conducted at one of the largest educational institutions in Saudi Arabia. The quantitative phase included 317 valid student responses analyzed using JASP. Principal Component Analysis confirmed the TAM-based factor structure, and repeated-measures ANOVA tested for differences across demographic variables such as gender, academic level, and years of English study. Findings indicated generally favorable perceptions of generative AI tools, with perceived ease of use rated higher than perceived usefulness. While no major gender-based differences were observed, modest distinctions emerged across academic levels and language study history. To contextualize these patterns, the qualitative phase involved 17 semi-structured interviews analyzed thematically using Quirkos. The qualitative findings provided deeper insight into how students incorporated AI tools into their learning routines across reading, writing, speaking, and listening tasks. Themes included increased learner autonomy, more personalized and immediate feedback, enhanced motivation, and flexible pacing. However, students also raised concerns about the accuracy of the tool, its insensitivity to context, the lack of critical reflection, and the risk of over-reliance on automated feedback. These nuanced findings were presented using bilingual excerpts (Arabic and English) to preserve cultural and linguistic authenticity. Together, the mixed-methods results suggest that generative AI tools offer both pedagogical opportunities and challenges. While learners recognize the value of these tools for improving task efficiency and engagement, they also navigate tensions between convenience and cognitive growth. Overuse of AI software, particularly for writing, may inhibit deeper learning and reduce confidence in independent language production. This dissertation contributes to the evolving body of research on AI in education by centering learner voices in a global context. It advances the theoretical application of TAM in EFL environments and underscores the need for culturally responsive, pedagogically guided integration of AI tools. The findings hold implications for curriculum designers, instructors, and edtech developers seeking to balance innovation with integrity in technology-enhanced language learning.