- From Acceptability to Legitimacy Construction
Abstract
Generative AI tools are reshaping academic writing. The central issue is not their use, but when AI-assisted text can still be recognised as legitimate scholarly work. This exploratory study examines how experienced academics evaluate legitimacy through three conditions: retention of human responsibility for core ideas, limitation of AI to supportive roles, and disclosure of its use. Data were collected from 25 participants through a questionnaire combining rating scales and open-ended responses. The findings show that legitimacy is conditional rather than binary. Participants accepted AI for drafting, rephrasing, and organising text, but expressed concern when it shaped arguments or interpretations. Across responses, three conditions consistently defined acceptable use: AI must support rather than replace intellectual work, authors must remain accountable for all claims, and AI involvement must be disclosed. Legitimacy, therefore, rests on ongoing professional judgment rather than fixed rules.
Keywords
AI-assisted Academic Writing, Legitimacy, Authorship, Accountability, Academic Governance, Disclosure, Higher Education
Introduction
G
enerative artificial intelligence has entered academic writing at a remarkable speed. The bigger question, however, is not simply how these tools are used, but whether the resulting text can still be seen as genuine scholarly work for which someone can be held accountable.
Academic writing has always been more than putting words on a page. It involves taking a position, defending claims with evidence, and standing behind one's arguments. When AI helps generate or reshape text, the central concern becomes whether authorship understood as responsibility and accountability remains intact.
Many current discussions focus on policies, detection tools, or outright misuse. While these matters are important, they do not fully capture how academics actually evaluate writing in daily practice during supervision, peer review, or manuscript preparation. In real academic settings, judgments about legitimacy are made through ongoing interpretation rather than rigid checklists.
This study therefore approaches AI-assisted writing as a question of legitimacy construction. It examines how experienced academics draw boundaries around acceptable use by looking at responsibility, the limitations placed on AI, and the role of disclosure.
The main research question guiding the work is:
How do academics construct and evaluate the legitimacy of AI-assisted academic writing in relation to responsibility, limitation of AI use, and disclosure in higher education contexts?
Literature Review
AI in Academic Writing in Higher Education
Generative artificial intelligence has become increasingly visible in academic writing practices, particularly through systems that assist with drafting, paraphrasing, summarising, and language refinement. These tools are often described as productivity aids that support early stages of writing, especially where clarity and organisation are concerned (Kasneci et al., 2023; Cotton et al., 2023). However, framing AI primarily in terms of utility risks obscuring the more fundamental issue it introduces: not what these systems can do, but how their use affects the conditions under which academic writing remains legitimate.
This distinction becomes clearer when academic writing is understood not as a technical skill but as a socially regulated practice. Hyland (2016) demonstrates that writing is central to the construction of identity and authority, requiring writers to position themselves credibly within disciplinary communities. Similarly, Lea and Street (1998) show that writing practices are embedded within institutional expectations that define what counts as valid knowledge. When AI systems contribute to textual production, these expectations are not removed but placed under pressure. The question is therefore not whether AI can produce acceptable text, but whether such text can still be recognised as the outcome of accountable authorship.
This pressure becomes more pronounced when considering the nature of AI-generated output. Large language models operate through probabilistic patterning rather than conceptual understanding, producing text that may appear coherent while lacking epistemic grounding (Bender et al., 2021; van Dis et al., 2023). As Mittelstadt et al. (2016) argue, such opacity introduces a problem of accountability, where the basis on which knowledge claims are generated becomes difficult to trace. In academic contexts, where legitimacy depends on the visibility of reasoning and evidence, this creates a structural tension between fluency and trust.
At the same time, broader discussions of AI in knowledge work have emphasised the need to clarify how responsibility is distributed between human and automated agents. Floridi et al. (2018) argue that ethical frameworks for artificial intelligence must address not only technical risk but the allocation of responsibility within hybrid systems. In academic writing, this issue becomes central because the legitimacy of a text depends not on its production alone, but on the ability to attribute responsibility for its claims. The integration of AI therefore raises a governance problem: how responsibility is maintained when aspects of writing are partially automated.
Taken together, these strands of research suggest that AI in academic writing cannot be understood solely in terms of capability or efficiency. The central issue is how its use interacts with the conditions that make academic writing recognisable as legitimate. This shifts the focus from tool use to the regulation of authorship and accountability.
Authorship, responsibility, and Academic Writing
Authorship has long functioned as a central mechanism through which academic work is regulated. To be recognised as an author is not only to contribute text, but to assume responsibility for the interpretation of evidence, the coherence of argument, and the validity of claims. As Biagioli (2012) demonstrates, authorship historically emerged as a system for assigning both credit and accountability within scientific communities. This dual function becomes unstable when textual production is distributed across human and non-human agents.
From a research ethics perspective, authorship is explicitly tied to accountability. Resnik and Shamoo (2011) emphasise that authorship entails the capacity to defend and take responsibility for published work. Similarly, Resnik and Elmore (2015) argue that research integrity depends on the ability to attribute responsibility clearly within collaborative processes. These frameworks assume that those who produce text are also those who can be held accountable for it an assumption that becomes difficult to sustain when AI systems generate or shape textual content.
This difficulty is compounded by the fact that academic writing has always involved forms of assistance. Editorial support, peer feedback, and collaboration are integral to scholarly work. However, these forms of assistance remain embedded within systems of accountability because all contributors are, in principle, identifiable and responsible. AI introduces a qualitatively different form of assistance: it can generate text without bearing any responsibility for its meaning or accuracy. This creates a gap between production and accountability that existing frameworks of authorship do not fully resolve.
Work in academic integrity studies reinforces this point. Macfarlane et al. (2014) and Bretag (2016) show that norms surrounding authorship and attribution are not merely technical but reflect broader assumptions about intellectual responsibility. Sutherland-Smith (2010) further demonstrates that these norms are context-dependent and subject to interpretation. The introduction of AI intensifies this interpretive dimension by introducing forms of textual production that do not map neatly onto existing categories of authorship or collaboration.
From a discourse perspective, authorship can be understood not simply as ownership of text, but as a position within systems of meaning-making. Asghar (2014) argues that authority in academic writing is constructed through language, argumentation, and engagement with prior work. This perspective shifts attention from who produces text to who remains accountable for its interpretation. In the context of AI-assisted writing, this distinction becomes critical. While AI systems can produce linguistically plausible text, they cannot assume responsibility for its meaning, which remains attached to the human author (Lund et al., 2023; Thorp, 2023).
These developments suggest that authorship is being reconfigured rather than replaced. The defining feature of authorship is no longer exclusive control over text production, but the retention of accountability within processes that may involve automated assistance. This reconfiguration raises a governance question: how the conditions of authorship are maintained when writing becomes a hybrid process.
Transparency and Disclosure in AI-Assisted Writing
In response to these challenges, transparency has emerged as a key mechanism for regulating AI use in academic writing. Institutional and publisher guidelines increasingly require disclosure of AI involvement, particularly where it contributes to the production of content (Ganjavi et al., 2024; UNESCO, 2023). The rationale is that making the writing process visible allows readers and reviewers to assess the extent of human involvement and responsibility.
However, transparency does not resolve the problem of legitimacy on its own. Eaton (2023) notes that while disclosure is widely endorsed, there is limited agreement on what constitutes sufficient transparency. This ambiguity reflects a broader issue: visibility does not automatically translate into accountability. A text may disclose AI involvement while still raising questions about the extent to which the author remains responsible for its content.
Empirical analyses of journal policies support this point. Ganjavi et al. (2024) show that while many journals now address AI use, there is little consensus on how disclosure should be defined or verified. Similarly, Crawford et al. (2023) find that institutional responses vary between prohibition, detection, and integration, indicating that governance frameworks remain unsettled. UNESCO (2023) attempts to address this by linking transparency to human oversight, emphasising that disclosure must be accompanied by clear expectations regarding responsibility and verification.
These findings suggest that transparency functions less as a solution than as part of a broader regulatory framework. It provides visibility, but the evaluation of that visibility depends on how responsibility is interpreted. As a result, disclosure operates as one condition among others through which legitimacy is assessed, rather than as a sufficient mechanism in itself.
Practitioner Judgments and the Construction of Legitimacy
Despite the expanding literature on AI in higher education, much of the existing work remains focused on student behaviour, academic misconduct, or institutional policy. Less attention has been given to how legitimacy is constructed through everyday academic practice. This gap is significant because academic norms are not defined solely through formal guidelines, but through the judgments enacted in activities such as writing, reviewing, supervision, and publication.
Perkins (2023) argues that in periods of technological change, practitioners play a central role in interpreting acceptable practice, particularly in domains where norms are embedded in practice rather than codified in rules. Similarly, Wisker and Robinson (2016) show that supervision functions as a key site where expectations around authorship and responsibility are negotiated. These contexts reveal that legitimacy is not applied mechanically, but constructed through ongoing evaluation.
Recent work has begun to recognise this shift. Cotton et al. (2023) argue that detection-based approaches to AI use are insufficient because they assume that acceptable practice can be predefined. Instead, they call for frameworks that articulate how judgments about responsibility and acceptable assistance are made. Dwivedi et al. (2023) further note that the pace of technological change has outstripped the development of stable guidelines, making practitioner judgment a central site for observing how norms are formed in real time.
The present study builds on this perspective by examining how legitimacy is constructed through reference to responsibility, limitation, and disclosure. Rather than treating these as isolated attitudes, it analyses them as conditions through which academic writing continues to be regulated under changing technological conditions. In doing so, it shifts the focus from what practitioners think to how the boundaries of legitimate academic writing are actively produced and maintained.
Method
Design
This was a small-scale exploratory qualitative study designed to understand the practical judgments academics make about AI-assisted writing. Rather than seeking large-scale generalisation, the goal was to identify the recurring conditions and reasoning patterns participants used when deciding whether a piece of writing remained legitimate.
Data came from an online questionnaire that included both Likert-scale items and open-ended questions. The open-ended parts formed the heart of the analysis, as they allowed participants to explain their thinking in their own words.
Participants
Twenty-five individuals working in higher education took part. Most had postgraduate qualifications and many years of experience in teaching, research supervision, or scholarly publishing. They came from various regions including South Asia, the Middle East, Europe, Central Asia, and Africa. The sample was not chosen to be statistically representative, but because these participants regularly make real-world decisions about authorship and academic quality in their professional roles. Participation was voluntary and completely anonymous.
Analysis
I first summarised the structured responses to get an overall picture. The main work, however, involved careful reading and iterative coding of the open-ended answers. I looked for the specific conditions participants mentioned when they accepted, questioned, or rejected AI use. Over time, three interconnected conditions stood out clearly: limitation of AI's role, retention of human responsibility, and the importance of disclosure. These were not counted as simple themes but treated as the practical criteria that shape legitimacy in academic settings.
Instrument
Data were collected through an online questionnaire designed to elicit both general orientations and more elaborated accounts of AI use in academic writing. The instrument included structured Likert-type items alongside open-ended questions.
The structured items addressed issues such as familiarity with generative AI, perceived risks, and expectations concerning disclosure and the role of AI in writing. These items were used to provide a descriptive overview of how participants positioned themselves in relation to the issue.
The open-ended questions formed the main basis of the interpretive analysis. Participants were invited to explain how they distinguished between defensible and problematic forms of AI-assisted writing, under what conditions AI use became questionable, and how responsibility for AI-mediated text should be understood. These responses were treated not simply as personal opinions, but as accounts through which the conditions of legitimacy were articulated.
Findings
Legitimacy Is Conditional Rather Than Binary
The findings show that academics do not judge AI-assisted writing in simple yes-or-no terms. Instead, they evaluate it through a set of conditions that determine how much AI involvement is acceptable before authorship begins to feel weakened. Familiarity with generative AI was very common in the responses. Twenty out of the 25 participants reported using these tools themselves in their academic or professional work. However, this familiarity did not mean they accepted everything without question. Rather, it served as the background against which participants made more careful, nuanced judgments.
What mattered most was not the mere presence of AI, but the extent of its role in the writing process. Participants viewed AI-assisted writing as acceptable when it remained clearly in a supportive position for example, helping with drafting, polishing language, or organising ideas. However, they became uneasy when AI appeared to cross into the realm of intellectual contribution, such as shaping core arguments or interpretations. This line between assistance and substitution came up again and again in the responses. AI use was generally tolerated as long as it stayed within clear limitations; it was questioned or rejected when it seemed to take over the author's own role in developing arguments or interpreting data.
Several participants illustrated this boundary quite clearly. For instance, many stated that AI was fine for "polishing language or organising ideas," but not acceptable if it generated "the main claims of a paper." These distinctions did not appear as rigid rules. Instead, they functioned as practical criteria that academics apply when deciding whether a piece of writing still feels legitimate. In short, the central question for most participants was not whether AI had been used at all, but whether its use remained limited enough to preserve the author's intellectual ownership of the work.
Responsibility as the Core Condition of Legitimacy
One of the strongest and most consistent themes across the responses was that responsibility cannot be transferred to AI. Participants repeatedly emphasised that the human author must remain fully accountable for the meaning, interpretation, and validity of the text, no matter how much AI assistance was involved. Concerns about becoming intellectually dependent on these tools were common, with 19 out of 25 participants expressing this worry. They saw AI use as problematic precisely when it reduced the author's personal engagement with reasoning and argumentation.
This concern was often expressed through a clear boundary related to interpretation. Several participants noted that AI-assisted writing starts to feel unacceptable when the tool moves beyond language and begins generating meaning or analytical insights. One typical comment captured this view well: "if AI starts generating the interpretation, the researcher is no longer fully engaged." Such statements suggest a shared understanding that true authorship is defined less by who produces the text and more by who remains actively engaged with the argument and takes ownership of it.
In this way, responsibility emerged as the central condition for legitimacy. Even when AI helped with drafting or structuring the text, participants insisted that accountability for claims, sources, and interpretations must stay firmly with the human author. Ultimately, the legitimacy of AI-assisted writing is not threatened by assistance itself, but by any situation in which that assistance begins to obscure or displace the author's responsibility.
Disclosure as a Mechanism of Legitimacy
Disclosure stood out as another important way in which legitimacy is made visible and open to evaluation. A clear majority of participants (18 out of 25) believed that any use of AI in academic writing should be disclosed. Importantly, they did not see disclosure merely as a bureaucratic requirement. Instead, they described it as a way to maintain accountability in the relationship between the author and the readers or reviewers.
By openly acknowledging AI involvement, authors allow others to assess whether responsibility has been properly maintained. In this sense, disclosure helps build and sustain trust. Comments such as "readers need to know how the text was produced" and "disclosure shows that the author is still taking responsibility" illustrate how closely participants linked transparency with accountability. Disclosure, therefore, serves as a practical mechanism that makes the conditions of legitimate authorship observable.
At the same time, the responses revealed that disclosure is not yet understood in exactly the same way by everyone. Some participants felt that any level of AI use should be acknowledged, while others drew distinctions between substantial contributions and minor language support. This variation suggests that while disclosure clearly functions as one of the key conditions for legitimacy, its exact boundaries are still being worked out in practice.
Divergence and the Ongoing Negotiation of Legitimacy
Although clear patterns emerged, the responses did not show complete agreement. The presence of variation indicates that the norms around AI-assisted writing are still developing rather than fully settled. For example, opinions on whether AI use constitutes academic dishonesty were split: 11 participants expressed concern, 10 remained neutral, and 4 disagreed. This spread suggests that AI use is not automatically seen by everyone as a threat to academic integrity. Similarly, participants' views on the quality of AI-generated text varied widely, showing that legitimacy is not judged on output quality alone.
More importantly, there was a noticeable difference in where participants drew the line between acceptable assistance and problematic substitution. While nearly everyone recognised this distinction, the exact point at which assistance becomes substitution remained open to interpretation and seemed to depend on context.
Taken together, these points of divergence point to an active process of negotiation. The three conditions highlighted in this study limitation of AI's role, retention of human responsibility, and disclosure serve as important shared reference points. However, how these conditions are applied still varies. In the end, legitimacy in AI-assisted academic writing appears not as a fixed property of the text, but as something that emerges from ongoing evaluative practices within the academic community.
Table 1 Selected Response Patterns on AI-Assisted Academic Writing (N = 25)
Statement | Agree | Neutral | Disagree |
AI contributions acceptable in academic writing | 17 | 6 | 2 |
AI use may lead to intellectual dependence | 19 | 3 | 3 |
AI-generated content should be disclosed | 18 | 6 | 1 |
Discussion
The findings are best understood not simply as opinions about whether AI-assisted writing is acceptable or not, but as evidence of how legitimacy is actively constructed within academic discourse. Participants did not merely state whether they approved or disapproved of AI use. Instead, they described the specific conditions under which writing could still be recognised as accountable scholarly work. In this light, the core issue is not the presence of AI itself, but how legitimacy can be maintained when the production of text is partially automated.
This perspective is important because much of the current debate treats AI use primarily as a regulatory problem, focusing on issues of compliance, detection tools, or outright prohibition. Such approaches tend to assume that acceptable practice can be clearly defined in advance and then enforced through rules and policies. However, the present findings point to a different reality. Legitimacy is not something imposed from outside through policy alone; it emerges through the evaluative judgments that academics make in their everyday practice. These judgments work by establishing the conditions under which a text can still be attributed to a responsible author.
Conditions as Mechanisms of Discursive Regulation
The three conditions that appeared repeatedly in the data limitation of AI involvement, retention of human responsibility, and disclosure serve as practical mechanisms through which legitimacy is regulated. These conditions do not function as rigid rules. Rather, they offer a flexible framework for assessing whether AI-assisted writing stays within the accepted boundaries of academic conduct.
From a discourse perspective, these conditions can be seen as boundary-setting devices. They help mark the point at which helpful assistance turns into problematic substitution and where authorship begins to feel weakened. In doing so, they shape not only what is permitted, but what can actually be recognised as legitimate academic writing. The focus on limitation reflects worries about how far AI should be allowed to intervene; the emphasis on responsibility highlights the need for authors to remain accountable for meaning and argument; and the stress on disclosure ensures that these relationships remain visible to others.
Taken together, these three conditions form a kind of distributed governance. Instead of depending on formal enforcement, they work through shared expectations that guide how authors, reviewers, and readers evaluate texts. Legitimacy, therefore, is maintained not through prohibition, but through the continuous negotiation of these conditions in day-to-day academic practice.
Reconfiguring Authorship Under AI Mediation
The findings also suggest a gradual reconfiguration of what authorship means in the presence of AI. Traditionally, authorship has been closely linked to full control over text production. AI complicates this view. What was surprising in the responses was that participants did not insist on purely human-generated text. Instead, they placed the strongest emphasis on the retention of responsibility as the true essence of authorship.
This shift moves authorship away from a purely production-based model toward an accountability-based one. What ultimately matters is not whether the author wrote every single word, but whether they remain answerable for the content. AI-assisted writing becomes problematic when this connection between the author and the text is obscured, rather than simply when AI is used.
From a governance point of view, this change helps academic writing stay stable even as technology evolves. Rather than trying to redefine authorship in technical terms, participants reaffirmed its fundamental role as a way of assigning responsibility. This indicates that authorship continues to serve as an important regulatory concept, even when writing processes become more distributed.
The Limitations of Policy-Centred Approaches
These findings also draw attention to the limitations of relying too heavily on policy-centred responses to AI in academic writing. Institutional guidelines usually address AI use in terms of what is permitted, restricted, or must be disclosed. While such policies offer useful structure, they do not fully explain how legitimacy is actually judged in real academic settings.
The variation in participants' judgments shows that acceptable use cannot be reduced to a simple set of fixed rules. The same type of AI assistance might be viewed differently depending on its relationship to responsibility and authorial control. This suggests that effective governance cannot depend only on formal policies. It must also take into account the interpretive practices through which academic norms are lived out in practice.
In this context, the conditions identified in the study provide a more flexible way of thinking about governance. They do not dictate exactly how AI should or should not be used, but instead offer criteria for evaluating its use. This approach moves the focus from mere compliance to informed judgment, and from rule enforcement to the ongoing maintenance of accountability.
Legitimacy as an Ongoing Negotiation
The lack of complete consensus among participants further indicates that legitimacy in AI-assisted writing is not yet fully settled. While there was broad agreement on the importance of responsibility and disclosure, the exact boundaries of acceptable use remained less clear. This variation suggests that academic norms around AI are currently in an active phase of negotiation.
From a discourse perspective, such negotiation is not a sign of weakness or instability, but a normal part of how norms develop over time. As new technologies change the conditions of writing, long-standing concepts such as authorship and responsibility are being reinterpreted rather than abandoned. The conditions highlighted in this study can therefore be seen as temporary stabilisations useful shared reference points that guide evaluation while remaining open to future adjustment.
Implications for Academic Governance
The analysis carries several implications for how AI-assisted writing should be governed in academic contexts. Rather than viewing AI as an external threat that needs strict control, the findings suggest that governance efforts should focus on preserving the conditions that keep responsibility visible and clearly attributable.
This means moving away from purely rule-based regulation toward more condition-based guidance. Policies that simply list permitted or prohibited uses are unlikely to capture the full complexity of real writing practices. In contrast, frameworks that emphasise responsibility, limitation, and transparency offer a more adaptable foundation for governance.
Such an approach also fits better with how academic work is actually evaluated in practice. Peer review, supervision, and publication already rely on judgments about an author's credibility and accountability. Integrating AI into these processes therefore does not require inventing entirely new rules, but rather clarifying how existing expectations apply under changing technological conditions.
Repositioning the AI Debate
More broadly, the findings suggest that current discussions about AI in academic writing may be missing the main point. A great deal of attention has been paid to detection methods, potential misuse, or the technical capabilities of these tools. While these issues are certainly important, they do not address the central question raised by this study: how legitimacy can be maintained when text production is partially automated.
By shifting attention from tool use to the construction of legitimacy, the study reframes AI not as a problem that must be solved, but as a development that makes existing academic norms more visible. The boundaries of acceptable writing have always existed; AI has simply brought them into sharper focus by testing their limitations.
Practical Implications for Higher Education
The conditions identified in this study have clear practical implications for academic practice in higher education. In supervision, they offer a useful basis for discussing AI-assisted writing in terms of responsibility rather than simple permission or prohibition. In manuscript preparation, they provide criteria for deciding when AI use remains consistent with expectations of authorship, especially regarding interpretation and argument development. At the institutional level, they suggest that guidance on AI should go beyond general policy statements and instead explain how limitation, responsibility, and disclosure should be interpreted in everyday practice. These implications do not set rigid rules but support more consistent and thoughtful evaluative judgment in academic settings where AI use continues to grow.
Conclusion
This study set out to explore how the legitimacy of AI-assisted academic writing is constructed and evaluated under conditions of partial automation. The findings demonstrate that legitimacy does not depend on whether AI is used or not, but on the specific conditions through which responsibility, limitation of AI involvement, and disclosure are upheld in academic practice.
Working with these responses made it clear that across the 25 responses, these judgments were guided by recurring expectations. AI tended to be viewed more favourably when its role stayed limited, when responsibility for interpretation and argument remained clearly with the human writer, and when its use was made transparent. When these conditions were weakened, participants were more inclined to see AI use as incompatible with proper authorship.
These patterns suggest that current academic responses to AI-assisted writing form part of an ongoing process of norm formation. Participants were not rejecting AI outright; rather, they were actively negotiating the terms under which its use could remain consistent with expectations of accountability. In this sense, authorship continues to matter, not because every word must be written by a human hand, but because responsibility for the text must stay visible and clearly attributable to the author.
The findings also indicate that questions about AI-assisted writing cannot be settled through broad permission or prohibition alone. What is needed is greater clarity regarding the conditions under which AI use can remain acceptable within academic practice. This includes clearer guidance on the acceptable limitations of assistance, the continued importance of human responsibility, and the role of disclosure in building and maintaining trust.
From this perspective, the integration of AI into academic writing represents less a fundamental disruption and more a process of adjustment. Existing norms are not being swept away; they are being reinterpreted in response to new forms of mediation. The real challenge for higher education, therefore, is not simply to regulate AI as a tool, but to ensure that the conditions of authorship and accountability remain meaningful and intelligible in contexts where writing is increasingly supported by automated systems.
Declaration on the Use of Generative AI
Generative AI tools (such as ChatGPT) were used exclusively for language refinement, grammar checking, and improving readability and flow after the author had completed the full draft, including data analysis, coding, and interpretation. All ideas, research design, data collection, qualitative analysis, findings, arguments, and conclusions are entirely the author's own. The author takes full responsibility for every part of the content and has carefully reviewed and approved the final manuscript.
References
- Asghar, J. (2014). Language power and ideology in commercial discourse: A prologue to critical discourse analysis for neophyte analysts. Academic Journal of Interdisciplinary Studies, 3(4), 225–232. https://doi.org/10.5901/ajis.2014.v3n4p225
- Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610–623). https://doi.org/10.1145/3442188.3445922
- Biagioli, M. (2012). Recycling texts or stealing time? Plagiarism, authorship, and credit in science. International Journal of Cultural Property, 19(3), 453–476. https://doi.org/10.1017/S0940739112000226
- Bretag, T. (Ed.). (2016). Handbook of academic integrity. Springer. https://doi.org/10.1007/978-981-287-098-8
- Cotton, D. R. E., Cotton, P. A., & Shipway, J. R. (2023). Chatting and cheating: Ensuring academic integrity in the era of ChatGPT. Innovations in Education and Teaching International, 61(2), 228–239. https://doi.org/10.1080/14703297.2023.2190148
- Crawford, J., Cowling, M., & Allen, K. A. (2023). Leadership is needed for ethical ChatGPT: Character, assessment, and learning using artificial intelligence (AI). Journal of University Teaching & Learning Practice, 20(3), 1–12. https://doi.org/10.53761/1.20.3.02
- Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja, M., Albanna, H., Albashrawi, M. A., Al-Busaidi, A. S., Balakrishnan, J., Barlette, Y., Basu, S., Bose, I., Brooks, L., Buhalis, D., Wright, R. (2023). “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. International Journal of Information Mana
- Eaton, S. E. (2023). Academic integrity in the age of artificial intelligence. In S. E. Eaton (Ed.), Handbook of academic integrity (2nd ed., pp. 1–15). Springer. https://doi.org/10.1007/978-981-287-079-7_1-1
- Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People—An ethical framework for a good AI society: Opportunities, risks, principles, and recommendations. Minds and Machines, 28(4), 689–707. https://doi.org/10.1007/s11023-018-9482-5
- Ganjavi, C., Eppler, M. B., Pekcan, A., Biederstedt, E., Abreu, A., Collins, G. S., & Camm, C. F. (2024). Publishers’ and journals’ instructions to authors on use of generative artificial intelligence in academic and scientific publishing: Bibliometric analysis. BMJ, 384, e077822. https://doi.org/10.1136/bmj-2023-077192
- Hyland, K. (2016). Teaching and researching writing (3rd ed.). Routledge.
- Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., … Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences, 103, 102274. https://doi.org/10.1016/j.lindif.2023.102274
- Lea, M. R., & Street, B. V. (1998). Student writing in higher education: An academic literacies approach. Studies in Higher Education, 23(2), 157–172. https://doi.org/10.1080/03075079812331380364
- Lund, B. D., Wang, T., & Carr, R. (2023). ChatGPT and a new academic reality: Artificial intelligence-written research papers and the ethics of the new academic landscape. Journal of Academic Librarianship, 49(2), 102111. https://doi.org/10.1016/j.acalib.2023.102111
- Macfarlane, B., Zhang, J., & Pun, A. (2014). Academic integrity: A review of the literature. Studies in Higher Education, 39(2), 339–358. https://doi.org/10.1080/03075079.2012.709495
- Mittelstadt, B. D., Allo, P., Taddeo, M., Wachter, S., & Floridi, L. (2016). The ethics of algorithms: Mapping the debate. Big Data & Society, 3(2), 1–21. https://doi.org/10.1177/2053951716679679
- Perkins, M. (2023). Academic integrity considerations of AI-generated text in higher education: A ChatGPT case study. International Journal for Educational Integrity, 19(1), 1–15. https://doi.org/10.1007/s40979-023-00117-3
- Resnik, D. B., & Elmore, S. A. (2015). Ensuring the Quality, Fairness, and Integrity of Journal Peer Review: A Possible Role of Editors. Science and Engineering Ethics, 22(1), 169-188. https://doi.org/10.1007/s11948-015-9625-5
- Resnik, D. B., & Shamoo, A. E. (2011). The Singapore Statement on Research Integrity. Accountability in Research, 18(2), 71-75. https://doi.org/10.1080/08989621.2011.557296
- Sutherland-Smith, W. (2010). Retribution, deterrence and reform: The dilemmas of plagiarism management in universities. Journal of Higher Education Policy and Management, 32(1), 5–16. https://doi.org/10.1080/13600800903440519
- Thorp, H. H. (2023). ChatGPT is fun, but not an author. Science, 379(6630), 313. https://doi.org/10.1126/science.adg7879
- UNESCO. (2023). Guidance for generative AI in education and research. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000386693
- Van Dis, E. A. M., Bollen, J., Zuidema, W., van Rooij, R., & Bockting, C. L. (2023). ChatGPT: Five priorities for research. Nature, 614(7947), 224–226. https://doi.org/10.1038/d41586-023-00288-7
- Wisker, G., & Robinson, G. (2016). Doctoral education and the development of researchers: A guide for supervisors and students. Palgrave Macmillan. https://doi.org/10.1057/9781137320766
Cite this article
-
APA : Asghar, J. (2026). Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education. Global Social Sciences Review, XI(I), 195-206. https://doi.org/10.31703/gssr.2026(XI-I).16
-
CHICAGO : Asghar, Jabreel. 2026. "Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education." Global Social Sciences Review, XI (I): 195-206 doi: 10.31703/gssr.2026(XI-I).16
-
HARVARD : ASGHAR, J. 2026. Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education. Global Social Sciences Review, XI, 195-206.
-
MHRA : Asghar, Jabreel. 2026. "Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education." Global Social Sciences Review, XI: 195-206
-
MLA : Asghar, Jabreel. "Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education." Global Social Sciences Review, XI.I (2026): 195-206 Print.
-
OXFORD : Asghar, Jabreel (2026), "Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education", Global Social Sciences Review, XI (I), 195-206
-
TURABIAN : Asghar, Jabreel. "Constructing Legitimacy in AI-Assisted Academic Writing: Responsibility, Limitation, and Disclosure in Higher Education." Global Social Sciences Review XI, no. I (2026): 195-206. https://doi.org/10.31703/gssr.2026(XI-I).16
