AI evidence extraction

Arabic Academic Writing in the Age of GenAI: A Narrative Inquiry into Students Revision, Agency, and Dependency

Matrokhim MATROKHIM · Journal of Narrative and Language Studies · 2026

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.59045/nalans.2026.95

Methodology & findings

Study design

Qualitative narrative inquiry design with semi-structured interviews, revision notes, and AI-assisted revision records.

Sample

N = 8, 3 groups

Primary method

Thematic narrative analysis with iterative movement between descriptive coding and interpretive memoing. Within-case narrative reconstruction followed by cross-case comparison and pattern identification. Data triangulation, member checking, and audit trail procedures were used to strengthen trustworthiness.

Main result

The study found that "GenAI functioned primarily as a revision scaffold rather than as a substitute for writing." Students used GenAI to refine grammar, improve lexical choice, strengthen coherence, and produce a more formal academic register in Arabic. Additionally, "students' uptake of GenAI suggestions was negotiated through acceptance, modification, and rejection depending on intended meaning, disciplinary scope, source fidelity, and rhetorical appropriateness," demonstrating that "agency therefore remained central to GenAI-assisted revision, but it appeared less as complete independence from technology than as evaluative control exercised within human-AI interaction."

Reports effect sizes.

Research paradigm

Interpretive/constructivist

Author conclusions

"Pedagogically, the study suggests that Arabic writing instruction should move beyond simple permission-versus-prohibition debates and instead cultivate critical uptake, feedback literacy, and responsible GenAI-supported revision. The central question is not only whether students use GenAI, but whether they remain intellectually accountable for the texts they submit." The authors further conclude that "the most significant pedagogical risk is not GenAI use itself, but the gradual displacement of students' own evaluative judgment."

Risk of bias

Recall bias - participants' retrospective accounts of revision decisions may be selective or inaccurate; Social desirability bias - participants may have presented themselves as more critically engaged than during actual revision; Selection bias - purposive sampling of 8 participants from one institution limits representativeness; Researcher interpretation bias - study relied on researcher's interpretive memoing and coding choices; Recall bias from retrospective accounts; Self-presentation bias (participants may present themselves as more critically engaged than actual behavior); Selection bias from purposive sampling; Single institutional context limits generalizability; Small sample size (n=8); Self-presentation bias in participant narratives; Potential overestimation of critical engagement during actual revision; Single institutional context limiting generalizability

Limitations

The authors acknowledge that "the study involved a small number of participants from a single undergraduate context and therefore should be interpreted as exploratory and context-bound rather than representative of Arabic academic writing more broadly." They further note that "the analysis was strengthened through interviews, revision notes, and AI-assisted revision records, it still relied partly on self-reported and retrospective accounts
This means that participants' narratives may have been affected by recall bias or by a tendency to overestimate their own." The study also "focused on narrated experiences of revising specific Arabic academic tasks rather than on experimentally measured writing improvement or real-time observation of all revision decisions."

Open questions raised

Underrepresentation of Arabic academic writing contexts in GenAI research
Limited understanding of how students experience the lived process of revising Arabic academic texts with GenAI
Need for more process-oriented accounts of how students navigate GenAI-assisted revision in Arabic
Insufficient attention to the negotiated nature of revision in AI-mediated environments
Gap between experimental feedback comparisons and students' actual revision experiences with GenAI
Limited research on how students experience lived processes of revising Arabic academic texts with GenAI, particularly how they accept, modify, question, or reject AI-generated suggestions

Extracted from: pdfAgreement 67%

Explore related topics

AI Manuscript Writing Academic Writing Qualitative