Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1464
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-15T18:13:26.381549Z
status: published
lens: trace
sector: education
headline: The AI Mirror: A Qualitative Study of Student Autonomy and Dependence in ESL Writing
dek: This article explores the role of Artificial Intelligence (AI) in English as a Second Language (ESL) writing, examining whether it fosters autonomy or promotes dependency. Utilizing a qualitative methodology, semi-structured interviews were conducted following a timed writing task with 40 intermediate-level ESL students across three higher education institutions in Oman. Thematic analysis reveals that AI's impact on student autonomy is highly contingent upon its pedagogical application. When utilized strategical…
gain_title: When ESL students use AI strategically for brainstorming and personalized feedback, it works as a scaffold that supports self-directed learning.
problem_title: When ESL students lean on AI to generate whole texts or passively fix errors, they offload thinking and become less independent writers.
trace_subject: AI use in ESL student writing and learner autonomy
gain_reading: When ESL students use AI strategically for brainstorming and personalized feedback, it works as a scaffold that supports self-directed learning.
gain_evidence: When utilized strategically for brainstorming and personalized feedback, AI acts as a scaffolding tool that fosters self-directed learning.
problem_reading: When ESL students lean on AI to generate whole texts or passively fix errors, they offload thinking and become less independent writers.
problem_evidence: when students rely on AI to generate complete texts or passively correct errors, they experience cognitive offloading, leading to decreased independence
quick_read: Researchers interviewed 40 intermediate-level ESL students at three higher education institutions in Oman after a timed writing task. They used thematic analysis to ask whether AI tools build learner autonomy or encourage dependence in writing.

The findings suggest the outcome depends on how students use AI rather than on the tool itself, which puts weight on teaching practice and digital literacy. The study is qualitative and based on a small sample from one country, so it cannot show how common each pattern is or how large the effects are.
limitation: The evidence comes from interviews with 40 intermediate-level ESL students at three institutions in Oman, so it reflects self-reported experiences in a limited population rather than measured outcomes across wider groups.
tag: Dual reading
key_points: The study used semi-structured interviews following a timed writing task. | Participants were 40 intermediate-level ESL students across three Omani higher education institutions. | Strategic use for brainstorming and personalized feedback was linked to self-directed learning. | Generating full texts or passively correcting errors was linked to cognitive offloading and reduced independence. | The authors urge educators to build digital literacy so students treat AI as an intellectual partner.
rundown: The design paired a timed writing task with follow-up semi-structured interviews, and the interview data were analyzed thematically. This captures students' own accounts of how AI affected their writing process.

The central theme is that AI's effect on autonomy is contingent on pedagogical application. Brainstorming and personalized feedback made AI a scaffold, whereas full-text generation and passive error correction led to dependence.

The authors recommend that educators cultivate digital literacy so students use AI as an intellectual partner rather than a cognitive replacement.
sources:
- peer_reviewed | International Journal of TESOL & Education | https://doi.org/10.54855/ijte.26641 | 2026-10-15
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

6207f6a8c9fc9aa0d8e474ccae68a207f469763f9af2961747648702cb28ecc7

previous

0000000000000000000000000000000000000000000000000000000000000000