Plagiarism, a key form of academic misconduct, is one of today’s leading challenges facing higher education institutions across the globe (Awasthi, 2019; Miranda-Rodríguez et al., 2024; Stuhmcke et al., 2016; Yu et al., 2018). Pickard (2006) found that 72% of faculty at a UK institution reported to have encountered plagiarism cases in their work in the last academic year (p. 224). Harper et al. (2019) found, after surveying over 1,000 faculty members in Australia, that about 68% of faculty have “encountered a written task they suspected was written by someone other than the student who submitted it” (p. 1860). In the context of the UAE, very little research shed light on the impact of plagiarism on faculty experiences in the UAE. For example, Aljanahi et al. (2024) found that contract cheating results in faculty feeling that their and their students’ rights were being violated (pp. 8–9). They also reported that faculty sensed a lack of support from their institutions (p. 12). While the term plagiarism has been problematized by many scholars (Eaton, 2023; Howard, 2000), Fishman (2009) argues that it occurs when an individual “uses words, ideas, or word products attributable to another identifiable person or source without attributing the work to the source from which it was obtained in a situation where there is a legitimate expectation of original authorship in order to obtain some benefit, credit, or gain which need not be monetary” (p. 5). Pecorari and Petrić (2014) argues that plagiarism has been viewed traditionally as a form of “moral transgression and a reflection of decay” (p. 271) and as “literary theft that needs to be punished” (p. 272). However, the authors contend that academic integrity policies addressing plagiarism should consider intentionality as a critical factor, as instances of textual attribution may stem from insufficient knowledge of citation practices rather than deliberate misconduct. This consideration is particularly salient for non-native English speakers who may face additional challenges in navigating the conventions of academic writing in a secondary language. Following Chan’s (2024) work, this paper refers to plagiarism that does not involve Gen-AI as “traditional plagiarism.”

The emergence and rapid proliferation of Gen-AI has taken academic institutions across the globe by surprise, intensifying the long-standing concerns about plagiarism and the integrity of the learning mission of these institutions (Alshurafat et al., 2024; Chan, 2024; Crompton and Burke, 2024; Karkoulian et al., 2024; Kim et al., 2025; Sánchez-Vera et al., 2024). Sánchez-Vera et al. (2024) surveyed faculty members across Spain and found that the top three risks associated with Gen-AI were deterioration of essential student skills, excessive dependence on technology, and plagiarism (p. 20). They also found that 75% of faculty members have encountered Gen-AI plagiarism in their institutions. These concerns reflect broader apprehensions about the impact of Gen-AI on educational outcomes and integrity. Hostetter et al. (2024) examined perceptions of Gen-AI in academic writing among 83 students and 82 faculty at a Midwestern liberal arts college. Results revealed significant detection challenges, with only 29% of faculty and 27% of students being able to correctly identify AI-generated content. A clear ethical hierarchy emerged: spell checkers and citation generators were deemed most ethical, followed by AI use for outlines or specific sections, with whole-paper generation considered least ethical. Notably, perceived ethicality increased significantly when scenarios included student verification, editing, and acknowledgment of AI use (average increase: 0.79 points on a 4-point scale). Rather than implementing blanket bans, the authors recommend open discussions about the appropriate use of Gen-AI and its impact on learning. These findings suggest educational institutions should develop adaptive and balanced policies acknowledging the varying impacts of different AI applications while emphasizing student critical engagement with AI-generated content.

Submission of AI-generated work and taking claim of it can be argued as a new form of cyber-pseudepigraphy which Page (2004) defines as “using the Internet to have another person write an academic essay or paper, without this authorship being acknowledged” (p. 429). This paper argues that the definition can still encompass Gen-AI by adding to the definition “or a Gen-AI engine” after “another person.” The author goes on to claim that cyber pseudepigraphy is more serious than traditional plagiarism as “[w]ith plagiarism, there can be some doubt that the writer has unconsciously borrowed ideas that he/she might have read before, while with pseudepigraphy, there is no doubt that the action is fraudulent” (p. 431). Another study by Kim et al. (2025) found that half of the student believed it was ethical and appropriate to use Gen-AI to develop responses to assignments. In the same vein, Longoni et al. (2023) demonstrate that individuals perceive plagiarizing AI-generated content as less unethical than plagiarizing human-generated content. They further contend that the real challenge with Large Language Models (such as ChatGPT) lies in their availability, user-friendly interfaces, and the difficulty in detecting their output (p. 3). Low detectability of Gen-AI plagiarism provides students with the opportunity to cheat as the risk of being caught is smaller than that in the case of traditional plagiarism, with the exception of contract cheating where detectability has proven to be challenging to educators (Walker and Townley, 2012, p. 31).

Chan (2024) coins the term “AI-giarism” to refer to Gen-AI plagiarism and defines it as “[t]he unethical practice of using artificial intelligence technologies, particularly generative language models, to generate content that is plagiarized either from original human-authored work or directly from AI-generated content, without appropriate acknowledgement of the original sources or AI’s contribution” (p. 3). Like many definitions of traditional plagiarism, this definition overlooks intent on the part of students or the nature of the contribution by AI. It also seems to disagree with the findings of her own study where faculty ranked Gen-AI practices from most to least serious.

Academic integrity policies are important instruments that “strategically guide the [institutional] management of plagiarism” (Gullifer and Tyson, 2014, p. 1204) and are argued to help deter students from engaging in plagiarism (Bretag and Mahmud, 2016; Levine and Pazdernik, 2018). Bretag and his associates’ (2011) seminal work provided a comprehensive typology of the core components of an effective academic policy instruments including access, approach, responsibility, detail, and support. Cullen and Murphy (2025) point to a scholarly consensus supporting the robustness of this framework (pp. 5–6). Bretag et al. (2011) argue that for an academic integrity policy to be effective it should first be accessible to its target users and employs a clear language. In a later work, Bretag and Mahmud (2016) recommended the use of clear policies that guide faculty with managing plagiarism starting from suspension of misconduct and ending with imposition of sanctions. Similarly, a report commissioned by the Australian Universities Teaching Committee recommended “highly visible [institutional] procedures for monitoring and detecting cheating” (James et al., 2002, p. 37). However, availability and visibility of academic integrity policies alone do not necessarily correlate with reduced plagiarism on the part of students or increased enforcement on the part of faculty as found by Gullifer and Tyson (2014). For this reason, Bretag et al. (2011) posit that an effective policy should follow an educative approach. This corroborates the findings of Brown and Howell (2001) who concluded that educative policies are more likely to be effective. Third, they argue that an effective policy should clearly define all the stakeholders involved in academic integrity. This aligns with Gullifer and Tyson (2014) which called for active engagement with stakeholders to maximize the benefits of a policy instrument (p. 1214). Fourth, they contend that an effective policy instrument provides adequate details, allowing users to navigate the different subtleties of plagiarism. This comports with Niraula (2024) who calls for comprehensive policies that cover the different aspects of plagiarism, especially Gen-AI. Finally, they point to the importance of an institutional support apparatus enabling the implementation of the policy.

Bretag et al. (2011) framework has been used by many scholars (e.g., Cullen and Murphy, 2025; Stoesz et al., 2019) to evaluate written policy documents. However, it has not been used to examine faculty perceptions of policy documents. As such, this paper extends the scope of this framework to cover not only written policy but also faculty perceptions of these policies.

The next section of this paper outlines the research questions this study aims to answer.

This paper aims to further our understanding of academic plagiarism in the era of Gen-AI by answering the following four questions:

The study was conducted at a higher education institution in the United Arab Emirates where academic integrity constitutes a core institutional value. The institution has implemented regulatory frameworks addressing both traditional and Gen-AI plagiarism. Traditional plagiarism governance is characterized by extensive, scenario-based policies with corresponding sanction gradations. The institution employs Turnitin as its primary plagiarism detection mechanism, with institutional guidelines providing explicit thresholds for similarity percentages and associated disciplinary measures. In contrast, generative AI governance remains predominantly abstract, lacking the granular scenario-based approach and detailed procedural guidance that characterizes traditional plagiarism policies. This regulatory asymmetry between traditional and emerging forms of academic misconduct provides the institutional context for examining faculty perceptions of academic integrity violations.

Faculty members evaluated five key dimensions of plagiarism policies: availability, visibility, clarity, adequacy, and effectiveness. Availability, visibility, and clarity correspond to the ‘access’ element in Bretag and her associates’ (2011) framework, while adequacy corresponds to ‘detail.’ ‘Approach,’ ‘support’ and ‘responsibility’ are elicited from interviews as shown in the next section.

Results indicate a significant disparity in perceived quality between traditional and Gen-AI plagiarism policies, with traditional policies receiving substantially more favorable evaluations across all dimensions. Traditional policies received favorable majority evaluations in availability, visibility, clarity, and adequacy, with effectiveness approaching a majority approval. In contrast, Gen-AI plagiarism policies achieved favorable majority ratings only in availability, with borderline majority assessments in visibility and clarity, and a significantly low scoring for effectiveness. Comparative analysis (Figure 1) reveals substantial disparities between the two guideline types, with particularly pronounced differences in visibility (a 35 percent point gap) and adequacy (a 25 percent point gap). These findings suggest that the investigated institution, similar to its peer institutions globally, appears to be navigating the complex process of developing and implementing comprehensive policies that address the unique challenges presented by Gen-AI plagiarism. These perceptions among faculty corroborate the conclusions of McDonald et al. (2025) who found that while a majority of Gen-AI policies adopted by US institutions (63%) encourage Gen-AI use, only 43% provided detailed guidance to faculty and students on how to manage Gen-AI plagiarism. On the other hand, the findings in the interviews are consistent with this quantitative outcome as faculty members generally agreed that the policies for traditional plagiarism are understood by both students and instructors. In contrast, they expressed a significant concern that current policies and detection tools fall short when addressing AI-generated content. The emergence of tools like ChatGPT has created uncertainties, as noted by F3 in his/her remark:

This observation underscores the technological-policy gap currently facing higher education institutions and suggests the need for more agile policy frameworks that can adapt to rapidly advancing AI capabilities. This finding is consistent with the conclusions drawn by Miron et al. (2021), who argued for a need to update academic integrity policies to account for the evolving higher education landscape.

Deeper statistical analysis of faculty perceptions revealed statistically significant differences across all five measured dimensions. For example, paired-samples t-tests demonstrated that traditional policies were rated significantly higher than Gen-AI policies on effectiveness [e.g., M = 3.44, SD = 1.09 vs. M = 2.89, SD = 1.11; t(65) = 4.76, p < 0.001, d = 0.59]. The magnitude of these differences, as indicated by Cohen’s d values, ranged from medium to large effect sizes. Notably, the standard deviations for Gen-AI policies were consistently higher across all dimensions, suggesting greater variability in faculty perceptions of these newer policies. This pattern of results indicates not only that traditional policies are perceived more favorably, but also that there is greater consensus, whereas perceptions of Gen-AI policies demonstrate less uniformity, potentially reflecting the emergent and evolving nature of institutional responses to Gen-AI technologies. Furthermore, the data reveals that a thin majority of faculty believed traditional plagiarism policies were effective, with an overwhelmingly larger proportion holding an opposing view for Gen-AI plagiarism policies. This empirical evidence reveals a notable paradox: despite the ostensible visibility, availability, and adequacy of institutional policies—particularly those governing traditional forms of plagiarism—faculty members consistently report significant concerns regarding the effectiveness of these policies in addressing plagiarism. This discrepancy underscores the need for a thorough reevaluation of the impact of these policies and subsequent improvement of the existing policies governing both forms of plagiarism.

The next section addresses the remaining elements of Bretag and her associates’ (2011) framework, namely approach, support and responsibility.

The interviews revealed that faculty consistently advocate for more developmental rather than punitive approaches to academic integrity violations. This is exemplified by F2 who expressed concerns that by following policy s/he would “be causing [the student] to lose his career, just for something silly he made.” Along the same lines, F5 also remarked:

Faculty responses suggest that the policies are written in punitive, legalistic terms that overlook integrating educative approaches. This policy approach is not unique to the investigated institution as found by Stoesz et al. (2019) and Cullen and Murphy (2025) who concluded that academic integrity policies in Canadian and the American institutions are largely punitive than educative.

On the other hand, faculty perceive while the policy points to stakeholders who are involved in the process (including committees, student sponsoring agencies, the students themselves, and faculty), the practice places excessive responsibility on faculty members in proving the incidence of plagiarism. This is exemplified by F10’s comment arguing that: “[t]he onus here is very high for the faculty to prove, and we often cannot get to that burden.” A similar concern was echoed by F11 who noted:

Finally, interviewees repeatedly pointed to an urgent need for more institutional support to navigate the challenges associated with plagiarism. The section entitled Barriers to Policy Effectiveness addresses this matter in detail.

A majority of faculty (62%) indicated in the survey that they encounter traditional plagiarism instances, with 76% indicating that Gen-AI has increased plagiarism among students. This finding agrees with Harper et al. (2019) who found that about 67% of Australian faculty encounter academic misconduct and Eaton et al. (2019) who claimed that 90% of students in Canada self-reported engagement in dishonesty behaviors. When participants were asked whether they report these incidents, a majority indicated they do, with more faculty indicating that they report traditional plagiarism cases than Gen-AI cases (66% vs. 62%). This also agrees with the same Australian study which found that 56% of faculty members indicated that they report these incidents. Despite this expressed attitude among the faculty, a review of institutional records reveals that an incommensurate number of plagiarism cases was reported for investigation over the past two academic years. The gap between attitude and behavior, on the one hand, and policy and practice, on the other, was articulated in the response to questions asking whether faculty allow students to re-submit their work based on percentage of plagiarism or type of plagiarism. In the response to these questions, 82 and 79% of faculty indicated that they do so, even though institutional policy requires them to report any case based on suspicion and regardless of type or percentage of plagiarism to help maintain a systematic record-keeping of incidents. This finding is consistent with de Maio et al. (2020) who concluded that “most academic staff do not appear to report” incidents in Australia (p. 103), Nelson (2021) who found that only two faculty members (out of 12) ever reported academic integrity incidents, and Eaton et al. (2019) who highlighted the disconnect between the prevalence of self-reported academic dishonesty and percentage of reported cases by faculty (p. 7). The survey included items that shed some light on the reasons for this response to academic incidents. First, faculty reported that plagiarism results in increasing their workload (60% for traditional plagiarism and 63% for Gen-AI plagiarism). This paper argues that under-reporting is a coping mechanism adopted by faculty members to manage their increased workload. This agrees with the findings of Sattler et al. (2017) who concluded that the more-time consuming the effort is, the less likely faculty are to report academic misconduct (p. 1138). It also corroborates the findings of Harper et al. (2019) who concluded that a quarter of faculty who do not report academic integrity incidents thought that the process was time consuming. Second, the perceived effectiveness of the policies (49% for traditional plagiarism policies vs. 28% for Gen-AI plagiarism) could be argued to affect the reportability of academic misconduct. This corroborates the findings of Sattler et al. (2017) who concluded that the lower the efficacy of detection, the lower the faculty are to use and report detection methods (p. 1129). Third, faculty members reported to have different interpretations of the policies regulating plagiarism, with much more faculty interpreting Gen-AI plagiarism differently (46% vs. 70%). This also agrees with Eaton et al. (2019) who noted in a review of the literature of academic plagiarism in Canada that there is a widening gap between policy and practice among faculty in Canadian Higher Education settings.

The interviews provided additional data on how faculty respond to plagiarism. Many participants reported using early intervention approaches to avoid reporting students. For example, F2 remarked:

This preventive approach reflects a discretionary decision to prioritize education over punishment aligning with the findings of Walker and White (2014) who concluded that “most participants maintained that educative strategies … were an effective preventative strategy [to plagiarism]” (p. 679) and Gottardello and Karabag (2022) who found that a majority of faculty participating in their study maintained that they view themselves as educators rather than protectors of academic integrity (pp. 532–3). Other participants expressed concerns about the proportionality of consequences relative to the offense. For example, F8 argued:

This concern about fairness and proportionality aligns Tummers and Bekkers (2014) who noted that faculty consider the impact of their decisions on their student lives, allowing them to make a difference in their lives when implementing the policy (p. 528). This approach is also consistent with Kier and Ives’ (2022) recommendation for policy review to ensure proportionality between academic integrity sanctions and the severity of student violations (p. 14). Others take matters into their own hands and reduce the student grade without reporting the incident. For instance, F5 remarked:

This finding aligns with the conclusions drawn by Coalter et al. (2007), who documented that faculty members frequently employ grade reduction as a punitive measure for plagiarism infractions, circumventing formal institutional reporting mechanisms. Concomitantly, this approach resonates with the ‘intimidator’ role faculty play, as documented by Gottardello and Karabag (2022), wherein faculty reported employing deterrence strategies and fear-inducing tactics to discourage academic dishonesty among students.

Following Amigud and Pell’s (2021, p. 929) call for institutions to endeavor to design plagiarism policies that are effective, this study examined faculty-identified impediments to academic integrity policy effectiveness through qualitative analysis of interview data. Seven distinct barrier categories were identified and quantified according to frequency of occurrence. Results indicate technological insufficiencies (n = 17, 100%) constitute the predominant concern, followed by and procedural limitations (n = 16, 94%), cultural impediments (n = 15, 88%), structural and systemic constraints (n = 14, 82%), faculty preparedness deficiencies (n = 13, 76%), accountability and enforcement gaps (n = 12, 70%), and student lack of awareness and preparedness (n = 8, 47%). These findings suggest effective policy implementation requires multidimensional interventions addressing both structural and cultural dimensions of the academic integrity ecosystem.

Technological insufficiencies emerged as the most frequently cited barrier (n = 17), with faculty reporting significant limitations in plagiarism detection, particularly regarding AI-facilitated misconduct. As F10 and F7 explicitly stated:

This finding is not surprising and comports with those of Lodge et al. (2023) and Alshurafat et al. (2024).

Procedural limitations (n = 16) also appeared repeatedly in the interviews when faculty spoke about sanctioning parameters, reporting protocols, and escalation hierarchies. One faculty member described, “a [team] member handled [plagiarism] very poorly and because of how strict the policy and procedures appeared… [he] … had no choice” (F6). These procedural limitations engender inconsistent enforcement practices, with faculty frequently resorting to extra-institutional resolution strategies rather than engaging with formalized reporting frameworks. As one respondent explained, “If you get 3 or 4 people together, you’ll get 3 or 4 different opinions. Some take a harder stance on plagiarism; others are more flexible” (F8). This finding is not unique to the UAE as found by Cullen and Murphy (2025) who concluded that academic integrity policies in the US are not consistently implemented and the argument of Brown and Howell (2001) who found that clearly written and educative policies are more likely to be effective.

Faculty interviewees also revealed that cultural impediments (n = 15) operate through institutional paradigms that privilege leniency over accountability, coupled with student value systems that prioritize assessment outcomes over learning. One faculty member observed:

F16 attributed this to a lack of student motivation, while F17 critiqued a “bargaining culture” where some students negotiate grades rather than reflect on misconduct. Faculty also contribute to this barrier through emotional reluctance to report, such as F2 who admitted avoiding penalties to prevent student dismissal. This cultural leniency normalizes non-compliance, as students perceive consequences as rare or mild. This finding agree with those of Tummers and Bekkers (2014) who contends that faculty consider the impact of their decisions on students’ lives when implementing the policy (p. 528).

Faculty also reported structural and systemic constraints (n = 14) as a driving factor creating environmental conditions that inadvertently incentivize academic misconduct while simultaneously discouraging rigorous enforcement. F2 observed that high academic workloads and systemic pressure act as barriers. F5 noted that plagiarism checks are labor-intensive and unrewarded, leading faculty to prioritize teaching over enforcement. This perception aligns with the responses to the survey where 60 and 63% of faculty perceive traditional and Gen-AI plagiarism increase their workload. This finding is also consistent with Harper et al. (2019) who found that a quarter of faculty who do not report academic integrity incidents attributed this behavior to the load associated with reporting. In this vein, F3 noted:

Additionally, some faculty claimed that policies often conflict with practical realities. For example, F14 noted that faculty bypass formal reporting due to perceived futility, while F17 highlighted inconsistent penalties (e.g., some students were expelled while others were allowed to resubmit). Structural misalignment between institutional rhetoric and practice further erodes trust, as “integrity” values remain aspirational without operational frameworks (F14 and F16). This systemic barrier can be argued to enable plagiarism and weaken policy enforcement.

Within the constrained environment, faculty and student training becomes increasingly important yet remains elusive. In this vein, faculty agreed that insufficient training and excessive workloads hinder consistent enforcement. Many faculty, like Interviewees 8 and 15, lacked familiarity with policies, relying on personal judgment. F2 specifically identified a lack of policy awareness and training as a critical barrier. Another, F1, noted that faculty may lack standardized tools or training to detect plagiarism reliably, creating loopholes in policy implementation. F9 stressed the need for interactive workshops to calibrate understanding of policy over passive training, while F14 criticized the absence of AI ethics programs. This finding agrees with those of Amigud and Pell (2021) who contend that policy communication, through training and awareness, is key to policy effectiveness.

Faculty also reported accountability and enforcement gaps as another reason for policy ineffectiveness. The burden of proof and fear of repercussions deter reporting. F2 emphasized the difficulty of proving Gen-AI plagiarism, while F4 noted faculty fear administrative backlash. Informal resolutions seemed to dominate, as F15 pointed out that his/her approach is typified by consulting with his/her program chair rather than filing reports. Inconsistent penalties—such as allowing resubmissions instead of reporting (F17)—undermine deterrence as argued by F4, F15, and F9. F11 highlighted ghostwriting’s evidentiary challenges, leaving faculty powerless without “smoking gun” proof. The finding here is consistent with many studies that tackled accountability gaps as the driving reason for the under-reportability of plagiarism. For example, Malak (2015) found establishing legally valid evidence of student engagement in plagiarism was one of the main reasons that reduced the likelihood of faculty reporting academic integrity incidents. Coalter et al. (2007) also found that 90% of faculty who did not report integrity incidents did so because of lack of proof (p. 7). Similarly, LeBrun’s (2023) findings revealed that a significant majority (87%) of participants reported difficulties in establishing sufficient evidence to meet the required burden of proof, particularly in cases involving Gen-AI plagiarism.

Student disengagement, while less frequently cited, also represents a significant barrier characterized by policies that inadequately address fundamental motivational deficits and misalignment between institutional expectations and student capacity. F1 identified a lack of student engagement with policies as a primary concern. Language barriers and inadequate institutional scaffolding exacerbate this. F7 and F6 noted that those whose pre-tertiary schooling is from rote-learning systems struggle with citation practices, leading to unintentional plagiarism. F6 and F8 described non-native speakers misusing AI to compensate for poor English. While less frequent, this theme underscores the need for proactive education to bridge gaps between secondary and tertiary expectations. LeBrun’s (2023) research corroborates this finding, demonstrating that a predominant proportion of faculty members identified cultural disparities and linguistic challenges as significant impediments to international students’ comprehension of and adherence to academic integrity protocols.

Section four of the survey included possible Gen-AI plagiarism scenarios and respondents were asked to rank their seriousness on a scale of 1 to 5. As shown in Table 1, the data shows that direct copying of AI-generated content (scoring 3.46/5) and using AI for plagiarism (3.11/5) are considered the most serious or unethical infractions, while using AI for grammar checking (1.85/5) and rephrasing one’s own content (1.98/5) are viewed as relatively acceptable practices. A strong correlation (0.78) between scenario seriousness and response rate suggests that more concerning scenarios elicit greater faculty engagement. The narrow spread of scores (SD = 0.52) indicates moderate consensus among faculty, with clear distinctions between problematic and acceptable AI uses. Faculty appear to differentiate based on the degree of student contribution: scenarios where AI replaces student work are viewed negatively, while those where AI augments student-generated content is more acceptable. Notably, simply acknowledging AI use does not significantly mitigate concerns when the underlying practice involves minimal student input. Concomitantly, the data shows that Gen-AI generated scenarios much less response rates, if compared with traditional plagiarism scenarios (next section). This paper argues that there are two main possible reasons for this observation. First, some faculty members are struggling to navigate the subtleties of Gen-AI plagiarism amid insufficient institutional guidelines and professional development opportunities, resulting in an inability to formulate definitive positions on these emerging scenarios. This is not surprising given the nascency of this development in higher education and corresponds with empirical evidence from multiple investigations examining faculty engagement with plagiarism detection, particularly concerning AI-generated content. Petricini et al. (2023) found that faculty were uncertain about the ethicality of using Gen-AI in higher education settings. Hostetter et al. (2024) and Kofinas et al. (2025) demonstrated that faculty were unable to reliably distinguish between student-authored and AI-generated submissions. Similarly, Freeman (2024) found that lack of policy specificity undermines the ability to clearly identify the appropriate use of Gen-AI. Second, the positioning of this inquiry within the terminal section of the survey instrument may have contributed to diminished response rates, potentially attributable to respondent fatigue—a methodological limitation warranting consideration in the interpretation of these findings.

The findings of this study offer a compelling, complementary faculty perspective that substantively corroborates the student-centered patterns identified by Chan (2024). The hierarchical evaluation of AI-assisted behaviors mirrors Chan’s findings with remarkable consistency; both datasets reveal that direct AI utilization (prompt input and submission) received higher ratings than more elaborate engagement strategies involving verification, editing, and synthesis. This parallel suggests that faculty and students share similar evaluative frameworks regarding AI tool usage in academic contexts, despite their different roles in the educational ecosystem. The close numerical alignment between our faculty-reported scores and Chan’s student-reported means (with differences ranging from just 0.01 to 0.41 across scenarios) indicates a surprising consensus between these two stakeholder groups. In this sense, the alignment between student and faculty attitudes corroborates the findings of Kim et al. (2025) who concluded that “students and faculty did not differ significantly in their attitudes toward Gen-AI in higher education” (p. 1).

However, a deeper statistical analysis reveals that faculty and student perspectives on AI-assisted work vary significantly in both severity assessments and consensus levels. Comparative analysis of faculty data versus student responses demonstrates that faculty consistently assign higher severity ratings to AI-assisted academic practices while exhibiting substantially greater response variability. For instance, regarding direct AI-generated content submission, faculty ratings (M = 4.37, SD = 7.23) contrast markedly with student evaluations (M = 3.44, SD = 1.18). This pattern persists across all measured scenarios, from simple grammar checking (faculty: M = 2.97, SD = 4.70; students: M = 2.09, SD = 1.29) to more complex AI integration practices.

The pronounced standard deviations in faculty responses (ranging from SD = 4.57 to SD = 7.47) compared to the relatively uniform student assessments (consistently SD = 1.13–1.29) constitute a critical finding with substantial implications for academic policy development. While students demonstrate cohesive understanding of AI-related academic integrity boundaries, faculty judgments reflect a heterogeneous evaluative framework, likely stemming from diverse disciplinary backgrounds, pedagogical philosophies, and varying degrees of technological familiarity.

Analysis of traditional plagiarism scenarios, Table 2, reveals a clear hierarchy of perceived violations, with intentional deception through ghostwriting (4.60) and submission of others’ work (4.38) viewed most severely, while technical citation errors (2.50) and unacknowledged language assistance (2.45) are considered less serious infractions. The data demonstrates that proper acknowledgment significantly reduces perceived severity across similar actions, as evidenced by the substantial difference between paraphrasing without acknowledgment (3.71) versus with acknowledgment (2.79). Notably, context matters significantly in translation scenarios, with direct translation of sources without attribution (4.13) viewed much more severely than using translation services to render one’s own work (2.43). Self-plagiarism through submitting identical work across courses (3.57) occupies a middle ground of concern, while unauthorized collaboration (3.64) is perceived as moderately serious. These findings suggest academic institutions should implement gradual sanctions aligned with violation severity, while focusing educational efforts on clarifying gray areas where student confusion may exist, particularly regarding translation services, collaboration boundaries, and proper acknowledgment practices.

A comparative analysis with Chan (2024) reveals significant discrepancies between faculty and student evaluations of academic integrity infractions. Faculty assessments demonstrate elevated mean severity ratings for direct plagiarism and contract cheating behaviors, with a particularly pronounced disparity in the evaluation of ghostwriting (faculty M = 4.60, SD = 0.85; students M = 3.99, SD = 1.27). This substantial difference suggests a fundamental difference in how these groups conceptualize the severity of outsourced academic work. Conversely, for textual appropriation involving paraphrasing without attribution, student perceptions (M = 3.95, SD = 1.12) indicate marginally higher severity ratings than faculty evaluations (M = 3.71, SD = 1.11).

The heterogeneity in standard deviations between these populations merits particular attention as it illuminates the degree of consensus within each group. Faculty responses to ghostwriting demonstrate notably lower variance (SD = 0.85), indicating substantial agreement among faculty. In contrast, student responses exhibit greater dispersion across multiple categories, suggesting more diverse interpretative frameworks regarding plagiarism.

The current faculty-focused study makes a significant contribution to the academic integrity literature by addressing critical scenarios not covered in Chan’s student-centered investigation. Specifically, it introduces and quantifies faculty perspectives on language assistance without acknowledgment (M = 2.45, SD = 1.39), translation services usage (M = 2.43, SD = 1.42), informal peer idea incorporation (M = 2.65, SD = 1.27), and citation formatting errors (M = 2.50, SD = 1.29)—all representing increasingly common academic practices in multilingual and collaborative learning environments that have been underexplored in previous research.

Faculty concerns encompass a granular spectrum of academic integrity issues, particularly partial or technical violations often overlooked in student conceptualizations. This measurement of “gray area” scenarios provides crucial insights into academic integrity’s evolving landscape amid digital tools and collaborative practices transforming traditional authorship concepts.