Writing back with AI - a phenomenological study of language decolonization among Arab bilingual learners

Oct 15, 2026

Upload by jayni

This transcendental phenomenological study examines how bilingual Arab writers experience language decolonization when using AI-generated tools for academic writing. A snowball-screened survey identified 32 respondents, from which nine undergraduate and postgraduate students in the United Arab Emirates were selected via criterion sampling. Interview data were analyzed with cluster analysis and sentiment analysis across six developmental stages, revealing tensions from dominant English reinforcement, cultural erasure, and linguistic mismatch. Critical engagement with AI feedback supported metalinguistic awareness and reassertion of bilingual identity, informing decolonial language theory and critical digital literacy.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages22

Reading Time55 min

AccessFree to download

Abstract

Introduction

Literature Review

Language Decolonization and AI

Contemporary Educational Technology, 2026, 18(2), ep651 ISSN: 1309-517X (Online) |
OPEN ACCESS Research Article
Writing back with AI: A phenomenological study of language
decolonization among Arab bilingual learners
Maha Alhabbash 1
0000-0002-2881-823X
Najah Al Mohammedi 1*
0000-0002-1331-2850
Hana Omar 1
0000-0001-9140-6740
Safa Alothali 2
0000-0002-2636-2697
1 United Arab Emirates University, Al Ain, UNITED ARAB EMIRATES
* Corresponding author: [201370306@uaeu.ac.ae](201370306@uaeu.ac.ae)
Citation: Alhabbash, M., Al Mohammedi, N., Omar, H., & Alothali, S. (2026). Writing back with AI: A phenomenological study of language decolonization among Arab bilingual learners. Contemporary Educational Technology, 18(2), Article ep651 .
[https://doi.org/10.30935/cedtech/18477](https://doi.org/10.30935/cedtech/18477)
ARTICLE INFO ABSTRACT | | | --- | --- | | Received: 22 Aug 2025
Accepted: 12 Jan 2026 | This transcendental phenomenological study explores how bilingual Arab writers experience language decolonization when using artificial intelligence (AI)-generated tools in academic writing. A screening survey was distributed using snowball sampling, yielding 32 respondents. From these, nine undergraduate and postgraduate students in the United Arab Emirates were selected through criterion sampling based on their responses, willingness, and availability to participate. Drawing on interview data, the study employed cluster analysis and sentiment analysis to trace the process of decolonization through six developmental stages: initial trust in AI authority, cultural and linguistic mismatch, discomfort with standardized norms, negotiation of voice, strategic adaptation, and advocacy for inclusive AI. Participants’ narratives revealed how AI tools often reinforced dominant English norms, creating tensions around authenticity, cultural erasure, and linguistic misalignment. These tensions prompted critical reflection and selective engagement with AI feedback, fostering greater metalinguistic awareness and a reassertion of bilingual identity. The findings contribute to decolonial language theory, critical applied linguistics, and critical digital literacy by grounding how algorithmic systems influence bilingual expression. This study proposes a grounded model of AI-mediated language decolonization, with implications for culturally inclusive AI design and linguistically responsive pedagogy in multilingual academic contexts.
Keywords: language decolonization, Arab bilinguals, college students, AI generative tools, academic writing | | INTRODUCTION | | The interaction between artificial intelligence (AI) and educational technology has progressed from rulebased systems to machine learning and neural networks. In language and literacy education, AI now supports automated essay scoring, grammar correction, and adaptive systems tailored to learners’ reading and writing levels (Caines et al., 2025) . A major advancement is the emergence of large language models (LLMs), AI systems trained on massive datasets to generate text that resembles human language (Omiye et al., 2024) . Based on transformer architectures that utilize self-attention mechanisms, LLMs can learn complex word associations and generalize beyond the training data to perform tasks such as text prediction, translation, and summarization (Shao et al., 2024). Progress in language modelling has been shaped by word embeddings, evolving from Word2Vec to contextual models like ELMo, which enable better syntactic and semantic representation, allowing for the development of massive models such as GPT-2, GLaM, and GPT-4 (Peters et al., 2018; Radford et al., 2019) . Tools like ChatGPT offer notable opportunities for language learning, especially in academic writing (Dizon, 2024; Godwin-Jones, 2024) . LLMs are now central to research in natural language processing (NLP) and AI, marking a shift in language learning technologies, pa

FAQs

What phenomenon does the study investigate about bilingual Arab writers using AI tools?

It investigates how bilingual Arab writers experience language decolonization while using AI-generated tools in academic writing.

How were participants selected for the study?

A screening survey used snowball sampling to yield 32 respondents, then nine undergraduate and postgraduate students in the United Arab Emirates were selected using criterion sampling based on their responses, willingness, and availability.

What analytical methods and stages were used to trace language decolonization?

Interview data were analyzed using cluster analysis and sentiment analysis, tracing the process across six developmental stages: initial trust in AI authority, cultural and linguistic mismatch, discomfort with standardized norms, negotiation of voice, strategic adaptation, and advocacy for inclusive AI.

Oct 15, 2026

Upload by jayni