The DOI matches its Crossref record; 38 of the 38 references it cites were found in a registry or an index; none of the 51 sources its claims cite is checked for support yet.
Negotiating Writer Identity in Second Language Writing with Generative AI Support
Open accessAbstract
As generative artificial intelligence (GenAI) becomes increasingly integrated into academic writing practices, questions surrounding second language (L2) writer identity require renewed attention. This study investigates how Bangladeshi university students—writing in English as a second language—construct and negotiate their authorial presence in GenAI-mediated environments. Drawing on sociocultural theories of identity, authorship, and language ownership, the research explores the paradoxical role of GenAI—as both a linguistic scaffold and a potential homogeniser of academic discourse. Using classroom-based inquiry, the research analyses student reflective writing from two cohorts (2021, n = 10; 2025, n = 20) alongside reflective interviews ( n = 9). The findings indicate that GenAI can support identity development when students engage critically with its outputs, strategically revise, and reflect on their own rhetorical intentions. However, heavy reliance on AI-generated language may undermine authorial voice and raise ethical concerns about originality and epistemic agency. The study argues for a reconceptualisation of L2 writer identity in AI-mediated writing as hybrid, reflexive, and digitally situated. Pedagogically, it highlights the importance of AI literacy, metacognitive awareness, and assessment practices that recognise the collaborative nature of human-AI writing.
Overview
CoChat Summary
TL;DR
This qualitative study compares 2021 human-authored and 2025 AI-assisted graduate writing by Bangladeshi L2 students, plus reflective interviews, to examine how generative AI shapes writer identity. It finds GenAI can scaffold confidence and disciplinary alignment but is associated with lexical convergence, rhetorical standardisation, and attenuated authorial presence, with students negotiating voice through selective revision.
Key findings
- AI-assisted texts had higher TTR (~0.42–0.48 for a 1,200-word sample) than non-AI texts (~0.20), but showed heavier reuse of formulaic academic phrasing (“research shows”, “it is important to note”, “studies indicate”) and lexical convergence.
- Stance/voice shifted: Table 1 reports first-person pronouns 18–22 vs 42–48 per 1,000 words, epistemic hedges 6–10 vs 26–30, assertive impersonal stance 34–40 vs 8–12, and future-oriented modals 10–12 vs 22–25 in AI-assisted vs human-authored writing, respectively.
- Rhetorically, AI-assisted writing was highly standardised (IMRaD-like, explicit signposting, template-driven), while non-AI writing was looser, genre-mixed, narrative/inductive, with variable transitions.
- Interviews (n=9) described AI as a scaffold for brainstorming, outlining, and language refinement and as a confidence booster, but also reported ambivalence about voice loss; students described selective revision and ethical positioning (e.g., AI as “tutor, not a ghostwriter”).
Methods
- Qualitative comparative design combining corpus analysis of two graduate writing corpora: 2021 human-authored (n=10, ~10,500 words; teaching philosophy statements/critical reflections, no AI) and 2025 AI-assisted (n=20, ~25,000 words; applied linguistics essays, AI use declarations, human-controlled).
- Corpus analysis: lexical diversity (TTR, word length), voice/stance markers normalised per 1,000 words, rhetorical move coding, and identity construction coding using Ivanić’s tripartite model.
- Reflective interviews: n=9 from 2025 cohort, purposive sampling, 45–50 min, thematic analysis (Braun & Clarke 2006); ethical approval 2025-SLASS-NSR:02.
Limitations
- The paper states the sample is small and from a specific institutional context, limiting generalisability, and reliance on self-reported experiences may introduce interpretive bias.
- The paper calls for larger, more diverse, longitudinal, and discipline-comparative research.
- Evident from the text: the 2021 vs 2025 comparison confounds AI use with cohort/time and task/genre differences (reflective teaching philosophy vs applied linguistics essays), and AI use was self-declared with varying degrees of assistance.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 10 Not valid · 43 Not checked · 246 Own results and interpretations
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“This pattern differs from the more standardised organisational structure observed in many of the 2025 AI-assisted papers, where essays often followed a clearer contrast 'Theory > Application > Critical Reflection' sequence.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: unclear — It is unclear why the paper says it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“No AI assistance was reported during the writing process.”
plainn = 10- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Cohesion also tends to be more variable.”
hedgedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Transitions frequently rely on simple or additive markers such as 'Also', 'Hence', or 'Later on', instead of the more sophisticated signposting like 'Conversely' or 'Ultimately' that appeared more frequently in the AI-assisted texts of 2025.”
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“As academic registers, students consistently use elevated verbs and nouns to describe relationships and effects, such as 'mitigate', 'cultivate', 'foster', 'facilitate', 'constitute', and 'underpin', besides strong academic collocations like 'empirical support', 'critical reflection', 'systematic study', and 'power asymmetries'.”
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
114 more claims
Create a free account to see all 299 claims and to verify the 34 references still open.
- Probably not a substantial claim. 115
- It may state two separate claims as one. 104
- The quote only partly states this claim. 5
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 4.5.4 Critical Awareness and Ethical Positioning
“'I check every sentence because I don't want to submit something I cannot explain myself.' (ST004).”
§ 4.5.4 Critical Awareness and Ethical Positioning
“'I check every sentence because I don't want to submit something I cannot explain myself.' (ST002).”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 4.5.4 Critical Awareness and Ethical Positioning
“'If the idea comes from AI, I don't feel comfortable submitting it unless I change it and understand it.'(ST003).”
§ 4.5.4 Critical Awareness and Ethical Positioning
“'Ifthe idea comes from AI, I don't feel comfortable submitting it unless I change it and understand it.'(ST006).”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 3.5.5 Interview Data Analysis
“Four core themes emerged: AI as a structural scaffold for outlining and organisation, voice negotiation reflecting awareness of voice loss or gain across disciplines, confidence and legitimacy tied to perceptions of academic competence, and ethical positioning regarding ownership and originality.”
§ Additional information
“Table A: Interview coding themes.Theme Description Focus AI as a scaffold AI used for structure/outlining Organization support AI as authority Reliance on AI phrasing Linguistic control Voice negotiation Awareness of voice loss/gain Identity tension Confidence & legitimacy Perceived academic competence Self-positioning Ethical positioning Ownership concerns Authorship boundaries”
Citations
Cites
Using thematic analysis in psychology
Braun, V., & Clarke, V. (2006) · Qualitative Research in Psychology · doi:10.1191/1478088706qp063oa · Cited by 194,414
ChatGPT for good? On opportunities and challenges of large language models for education
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., et al. (2023) · Learning and Individual Differences · doi:10.1016/j.lindif.2023.102274 · Cited by 7,326 · cited 3×
“…have demonstrated the potential of generative AI to sca ff old L2 writing through feedback, idea generation, and linguistic support (Yan 2023; Lee 2023; Kasneci et al. 2023), much of this research remains exploratory and is largely situated in Global North educational contexts (Dwivedi et al. 2023).”
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445922 · Cited by 6,988 · cited 3×
“Much of the scholarly discourse on generative AI in education has emerged from institutions in the Global North, particularly North America and Europe (Holmes et al. 2019; Bender et al. 2021; Dwivedi et al. 2023; Kasneci et al. 2023).”
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja, M., et al. (2023) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2023.102642 · Cited by 4,584
The Darker Side of Western Modernity: Global Futures, Decolonial Options
Mignolo, W. D. (2011) · Duke University Press eBooks · doi:10.1215/9780822394501 · Cited by 2,545
Pennycook, A. (2010) · doi:10.4324/9780203846223 · Cited by 1,856
Identity and Language Learning: Extending the Conversation
Norton, B. (2013) · Multilingual Matters · doi:10.21832/9781783090563 · Cited by 1,776
Writing and Identity: The discoursal construction of identity in academic writing
Ivanič, R. (1998) · Studies in Written Language and Literacy · doi:10.1075/swll.5 · Cited by 1,043
(2019) · Open Research Online (The Open University) · Cited by 1,004
McCarthy, P. M., & Jarvis, S. (2010) · Behavior Research Methods · doi:10.3758/brm.42.2.381 · Cited by 905
Research works printed but not linked: 5
- Lee, Jihyun. 2023. “Using ChatGPT to Support ESL Students’ Academic Writing: Opportunities and Challenges.” TESOL Journal 14 (3): e712.
- Lehman, S. 2020. “Academic Writing, Identity, and Reflexivity in Higher Education.” Journal of Academic Writing 10 (1): 1–15.
- Phillipson, Robert. 1992. Linguistic Imperialism. Oxford: Oxford University Press.
- Quijano, Aníbal. 2000. “Coloniality of Power, Eurocentrism, and Latin America.” Nepantla: Views from South 1 (3): 533–80. https://www.decolonialtranslation.com/english/quijano-coloniality-of-power.pdf.
- Reimers, F., Z. Azim, M. Palomo, and C. Thony, eds. 2024. Artificial Intelligence and Education in the Global South, Cham: Springer.
Cited by
Voice Recovery in AI-Mediated Writing: Rethinking Writer Identity in the GenAI Era
McKinley, J. (2026) · Digital studies in language and literature · doi:10.1515/dsll-2026-3014 · Cited by 3
ТВОРЧЕСКОЕ МЫШЛЕНИЕ В LLM-ОПОСРЕДОВАННОМ АКАДЕМИЧЕСКОМ ПИСЬМЕ: КРИТИЧЕСКИЙ ОБЗОР
Горюнова, Н. Б. (2026) · Ученые записки Института психологи РАН · doi:10.38098/proceedings_2026_06_03_08
Daud, S., & Ghosh, S. (2026) · Proceedings of the AAAI/ACM Conference on AI Ethics and Society · doi:10.1609/aies.v9i1.43179
Metadata
- Authors
- Rahnuma, N.
- Affiliations
- Independent University, Bangladesh (BD)
- Corresponding author
- Rahnuma, Naureen
- ORCID
- Rahnuma, Naureen
- Venue
- Digital Studies in Language and Literature
- Publisher
- Walter de Gruyter GmbH
- Published
- 2026-05-07
- Volume
- 3
- Issue
- 2
- Pages
- 246-283
- ISSN
- 2943-0607 (ISSN-L)
- OpenAlex
- W7160418224
- Cited by
- 3
- References
- 35
- Citations by year
- 2026: 3
- FWCI
- 16.61
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Digital studies in language and literature (publishedVersion, cc-by)
- Topics
- Social Sciences › Linguistics and Language › Discourse Analysis in Language Studies
- Social Sciences › Arts and Humanities › Language and Linguistics › EFL/ESL Teaching and Learning
- Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
- Keywords
- authorial identity, second language writing, generative AI, authorial presence, authorship, language ownership, reflective writing, originality, epistemic agency, AI literacy, metacognitive awareness, human-AI collaboration, English as a second language
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-11-18, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-11-18 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Negotiating Writer Identity in Second Language Writing with Generative AI Support. https://app.cochat.ai/papers/10.1515/dsll-2026-0009
- The paper
- Rahnuma, N. (2026). Negotiating Writer Identity in Second Language Writing with Generative AI Support. Digital Studies in Language and Literature. https://doi.org/10.1515/dsll-2026-0009
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.