Introduction
The 21st-century skills put significant importance on digitization and digital learning. Online and distance learning (ODL) and massive open online courses (MOOCs) have remained topics of much discussion for the last two decades. Through the continual evolution of lesson transactions and learning processes, themes like emerging technology marked its presence in the literature. MOOCs play a transformative role in blended learning models, with a focus on their alignment with academic performance and preferences for streamlined course structures. Altınay et al. (2024) emphasize the influence of faculty guidance alongside the need for tailored MOOCs designs, equitable access, and professional development to enhance learning outcomes and meet diverse student needs in the digital age.
The constant demand and the altering context of society drive the technology service providers to keep on bringing changes to their products or services. Such a technological evolution and revolution formed the genesis of digital innovation. If we are not wrong, the recent COVID-19 pandemic might have both catalyzed its momentum and mechanism.
Trailed by the notion of continually changing mechanisms, our pursuit of task automation introduced us to a relatively updated concept concerning artificial intelligence (AI). In a generalized form, AI may be described as a phenomenon where machines do the job of humans. The recent advent of generative artificial intelligence (GAI) has successfully overcome limitations. It presents a continuum of interaction between humans and computers. Additionally, GAI can produce human-like responses through texts, graphics, videos, etc. Multiple GAI tools are available to the community for specific purposes. Bozkurt et al. (2024) critically examine the integration of GAI in higher education, emphasizing its potential to personalize learning and enhance efficiency while cautioning against inherent biases and risks to equity, creativity, and human agency. They call for evidence-based research and conscious decision-making to ensure that GAI supports authentic learning experiences and upholds ethical educational responsibility.
The data mining approach of the generative and conversational agent (GAI tools here) provoked debate and dispersed different sections of society. While medical and health education (Newton & Dimopoulos-Bick, 2024; Raza et al., 2024; Small et al., 2024) have efficiently embraced this technology, the higher education sector remains uncertain. Most higher educational institutions (HEIs) have welcomed GAI with a progressive outlook (Kong et al., 2024; Lodge et al., 2023; Simms, 2024), whereas a few prefer to take an antagonistic stand (Michel-Villarreal et al., 2023).
Given this paradox, the key question is what factors fuel this debate. One significant factor is GAI’s openness and flexibility. When misused, GAI’s evolving nature can lead to a scenario where education is artificial rather than intelligent. The perceived anthropomorphic challenge necessitates a thorough examination of the existing literature.
Literature Review
Digital Innovation through GAI
In pursuit of GAI model development for the provision of text and data mining (TDM), Dermawan (2024) presented a critical view on what the European Union (EU) member states could learn from the Japanese “non-enjoyment” purposes. It was argued that a flexible TDM built upon the German notion of non enjoyment purposes could become an opening clause to foster innovation and creativity in the age of GAI. An opening clause allowing TDM with “non-enjoyment” purposes could be permissible under a three-step test. It further suggested that if there is no political will to safeguard “the right to read should be the right to mine” and development of a legislative environment that will advocate for 72 hours of response if technological protection measures (TPMs) are preventing TDM, and Robot Exclusion Standard as a warning when TDM is not allowed on a website.
Shifting our attention from the regional or national level towards the institutional setting, Holmström and Carroll (2024) described, “How organizations can innovate with GAI?” To what extent can the launch of ChatGPT innovate organizational practices? Will the innovation disrupt the system or be manageable? Two key dimensions, automation/augmentation, and four generic innovation strategies: traditional tools, basic automation, automated assistance, and assisted augmentation constituted the framework of the discussion. The strategies essentially differ not only in automation and augmentation for innovation but also in terms of risks and challenges faced in the process and available tactics for managing the process. They suggested that practitioners could harness ChatGPT effectively by aligning their innovation objectives with the appropriate strategy, whether enhancing creative processes or streamlining operational efficiency, thereby navigating the complexities of innovation with a more structured and strategic approach.
Ideating new product design sometimes appears to be a challenging task. Joosten et al. (2024) draw attention to even finer details. In the attempt, researchers compared the ideation quality of human beings and GAI. The blind expert evaluation showed that AI-generated ideas score significantly higher in novelty and customer benefit, while their feasibility scores are similar to those of human ideas. The findings suggest that, under certain circumstances, companies can benefit from integrating GAI into their traditional idea-generation processes.
The realm of digital innovation through GAI is not limited to the above discussion; instead, it touches upon other peripheries. To mention a few, areas like academic integrity (Duane, 2024), AI chatbots (Bridgelall, 2024), AI literacy (Sengewald & Tremmel, 2024), augmented reality (Shen et al., 2024), ChatGPT (Chiarello et al., 2024; Foung et al., 2024; Gupta & Yang, 2024; Ritala et al., 2024), customizable learning (Pesovski et al., 2024), ethical dilemmas (Singh et al., 2024), health science (Newton & Dimopoulos-Bick, 2024; Raza et al., 2024; Small et al., 2024), higher education (Rudolph et al., 2024), nutrition education (Barker et al., 2024), school education (Sattelmaier & Pawlowski, 2023), sustainable development (Nyhan & Marshalll, 2024), and, so on. The literature indicates the scope of innovation areas associated with GAI. The content and context of GAI research are evolving constantly. When no unified consensus is reached regarding the innovation results, perhaps the situation will lead to a diffused state of affairs.
Diffusion of Innovation Associating GAI
“What drives European organizations to invest in GAI, and what challenges do they face in 2023–2024?” The findings indicate that European organizations are primarily motivated to adopt GAI to enhance competitive advantage, grow market share, and foster innovation in products and processes. At the same time, the most significant challenges are ethical concerns, data privacy issues, and the absence of clear regulatory frameworks. The study concludes that a balanced approach, combining innovation with responsibility, is key to unlocking the full potential of GAI while mitigating risks (Orozco & Welin, 2024).
Ghanaian academics come to the scenario to investigate why there is a slow pace of digital transformation in education in Ghana using ChatGPT. It was found that most of the authors of publications about ChatGPT were not from the African continent or were affiliated with international institutions. A thematic analysis of interview data involving 34 academics in Ghana about ChatGPT revealed that most academics had limited knowledge about ChatGPT and AI-powered chatbots. There were mixed views about the relevance and usefulness of ChatGPT in work-related tasks (Adarkwah et al., 2023).
A study conducted in the United Arab Emirates (UAE) examined factors such as technology and employee influence on the adoption of AI in e-innovative projects. The survey included 1,037 government employees. The analysis revealed that technology, as an external factor, significantly and positively contributed to adopting AI e-innovation technology. Further analysis indicated that employee-related internal factors directly influenced the adoption of AI e-innovation technology (Hirzallah & Alshurideh, 2023). Other nationwide studies were centered on the organizational processes for adopting breakthrough technology in Japanese firms (Hoshino & Hirao, 2024), AI adoption in Chinese higher education (Huang et al., 2024), South African higher education (Khoalenyane & Ajani, 2024).
Multiple isolated reports appeared, covering areas such as GAI adoption in the classroom (Ghimire & Edwards, 2024), ChatGPT adoption among management students (Abdalla et al., 2024), higher education (Ivanov et al., 2024), morality and ChatGPT usage (Urdan & Marson, 2024), responsible use of AI in health science (Al-Dhaen et al., 2023; Ball et al., 2024), teacher education (Nyaaba & Zhai, 2024), and many more. de Bem Machado and Sharma (2024) conducted a bibliometric review to explore the integration of AI in higher education, analyzing 1,278 works authored by researchers from 91 countries. The study highlights various applications, including intelligent tutoring systems, adaptive platforms, and chatbots, which enhance personalized and inclusive learning. It emphasizes AI’s role in transforming education by fostering digital pedagogy, addressing challenges like self-regulated learning (SRL) in online settings, and enabling personalized learning trajectories. Recommendations include adopting AI-driven tools to improve teaching methods and education governance.
The examples above signified the possible usage areas of GAI in diffused form. To what extent can such non-human entities (technological innovation through GAI here) influence the habitual behavior of human civilization? In search of an answer to this question, our subsequent discussion channelizes us towards examining the literature of thing ethnography, positioning GAI at the core.
Thing Ethnography in the Realm of GAI
The public hegemony of the things world concerning central objects and perceived asymmetry in the material culture system was portrayed by Lucas and Robb (2021). The study accommodates Neolithic European and early modern Iceland culture into their perspectives:
Material culture forms a relational system of distributed reality – a thingworld. But how do we get beyond simply saying that all material culture is meaningful and entangled to understanding the internal structure of such systems? Is it a flat terrain among co-equal things? Or are some objects more important than others, as we might intuitively suppose? And if so, why? This article presents an initial discussion of the problem. Using vignettes from two thingworlds – one from early modern Iceland, one from Neolithic Europe– the authors discuss what were the central material things in each, and for what reasons. This suggests that objects may be systemically central in different ways, for instance, things which connect and mediate relationships of different kinds, things which are non-substitutable, and things which span multiple roles and contexts. (p. 1)
Huang et al. (2021) appeared with a new interactive tool named “Thing Constellation Visualizer” (ThingCV) for visualizing the social network of objects based on their co-occurrence as computed from a large collection of photos. ThingCV enables perspective-changing design explorations over the network of objects with scalable links. The results of eight participants showed that designers were actively engaged in identifying interesting objects and their associated clusters of related objects. The designers projected social qualities onto the identified objects and their communities. Furthermore, the designers changed their perspectives to revisit familiar contexts and to generate new insights through the exploration process.
A time when more debate is going on regarding the usability of GAI in educational exercises, Michel-Villarreal et al. (2023) captured the perception of ChatGPT on the challenges and opportunities it presents in higher education. Positioning the thing ethnography approach the research findings emphasize the urgent need for clear policies, guidelines, and frameworks to responsibly integrate ChatGPT in higher education. It also highlights the need for empirical research to understand user experiences and perceptions.
Investigations were conducted associating “thing ethnography” and “incremental prompting” to examine the applicability of GAI in experiential learning for authentic assessment of higher education. The study’s findings include the GAI’s capacity to contribute to reflective thinking, hands-on learning experiences and augmenting authentic assessment, enhancing teaching-learning practices, and cultivating higher-order skills. The researchers further emphasized the importance of responsible usage of GAI for educational purposes (Salinas-Navarro et al., 2024a–b).
The body literature concerned is further enriched by accommodating ideas on doing the right thing (Webb, 2014), entangled ethnography (Murray-Rust et al., 2019), ethnographic possibilities (Michalko, 1980), more-than-human-design and AI (Nicenboim et al., 2020), Morse things (Wakkary et al., 2017), the importance of “waiting field” (Mannay & Morgan, 2015), things as co-ethnographers (Giaccardi et al., 2016), and many more to mention.
The literature exemplifies the possibility of innovation through the application of GAI in different fields of society (vide digital innovation through GAI and diffusion of innovation associating with GAI). The diversified study findings accommodate data from academic and non-academic spheres of the community. The health science (Newton & Dimopoulos-Bick, 2024; Raza et al., 2024; Small et al., 2024) and engineering sector (Russo, 2024; Sundberg & Holmstorm, 2024) appear to have a positive outlook to embrace the technological movement. The higher education community (de Bem Machado & Sharma, 2024; Huang et al., 2024; Ivanov et al., 2024; Nyaaba & Zhai, 2024; Rudolph et al., 2024) is broadly diffused concerning the opportunities and challenges raised by GAI (Adarkwah et al., 2023; Bozkurt et al., 2024; Michel-Villarreal et al., 2023; Salinas-Navarro et al., 2024a–b). The debate extends to other conflicting issues such as organic vs. synthetic learning (similar to the nuance of Holmström & Carroll, 2024; Joosten et al., 2024; Shen et al., 2024), data saturation vs. data malignancy (trailing the notion of Dermawan, 2024), and ethical vs. unethical use of GAI (Singh et al., 2024). Such ongoing discourse necessitates further research associating GAI and (higher) education. Literature review shows that several isolated studies have been reported on similar concerns like academic integrity (Duane, 2024), ethical dilemmas (Singh et al., 2024), morality and ChatGPT usage (Urdan & Marson, 2024), responsible use of AI (Al-Dhaen et al., 2023; Ball et al., 2024). In such a context, hardly any collective research is noticed that was solemnized on addressing academic dishonesty in higher education, especially in the age of GAI. Such a knowledge gap in the existing literature necessitates the requirement of the present inquiry. Therefore, the current study focuses specifically on the ethical considerations (Orozco & Welin, 2024) of GAI usage towards addressing academic dishonesty in higher education.
Methods
Study Design
This study is focused on higher education (like Ivanov et al., 2024; Rudolph et al., 2024) and the theoretical underpinnings of digital innovation (e.g. Rudolph et al., 2024), diffusion of innovation (e.g. Ivanov et al., 2024) and thing ethnography (e.g. Michel-Villarreal et al., 2023). According to the literature review, we are now in a diffused condition when it comes to using GAI to achieve digital innovation in higher education. However, the media that appears here is within the field of thing ethnography. As a result, we structure the research design according to the theories. The systematic literature review (SLR) is articulated qualitatively using a thematic analysis approach (Bannister et al., 2023). Javadi and Zarea (2016), and Morgan (2022) discussed the concept of theme analysis and provided an agreement on its limitations. However, due to the technique’s methodological adaptability, the researchers in this study chose to apply it for a thorough examination of the research problem. The aim is to examine the viability of academic misconduct associated with using GAI in higher education. Six specified research questions (RQs) are framed to guide the research direction. The RQs are presented below:
RQ1. What is the viability of academic integrity in higher education in the age of GAI?
RQ2. What could be the impact of GAI on plagiarism and cheating in higher education?
RQ3. How far can GAI influence existing curriculum and assessment practices to inhibit academic dishonesty in higher education?
RQ4. What could be the individual, institutional, and technical concerns in accommodating GAI in higher education?
RQ5. What are the ethical and legitimate concerns coming into the picture while attempting to locate GAI in higher education?
RQ6. What does the existing body of literature convey on the GAI policy matters towards tackling academic dishonesty in higher education?
Data Collection
Numerous studies have explored academic dishonesty, academic integrity, and related issues in the context of GAI. This SLR is constituted on these existing reports. Data collection was performed on 15th May 2024 using the largest academic database, Scopus (like Sozon et al., 2024a). Specific search terms were employed to filter titles, abstracts, and keywords within the literature published between 2022 and 2024. In the study, 2022 is considered the initial year of data search because in this year ChatGPT made its debut.
A robust search technique was followed using search terms “Academic Dishonesty”, “Generative Artificial Intelligence”, “generative ai” OR “Academic Misconduct”, “Generative Artificial Intelligence”, “generative ai” OR “Academic Integrity”, “Generative Artificial Intelligence”, “generative ai”. An additional inclusion criterion of English-language publications yielded a database of 23 documents, including journals, book chapters, and conference proceedings. The perceived association with the current difficulty informed the reasoning behind the search terms chosen and the selection of 23 documents in this investigation. Additionally, the documents included publications from multiple fields, such as chemical education, higher education, management education, and medical education. Unlike other studies (Bukar et al., 2024; Wang et al., 2024), this research retains conference publications due to their global acceptability and quality control as indexed by Scopus. Data inclusion/exclusion criteria are summarized in the following Table 1.
Table 1
Data inclusion/exclusion criteria.
| VARIABLE | INCLUSION CRITERIA | EXCLUSION CRITERIA |
|---|---|---|
| Database | Scopus | Other than Scopus. |
| Year span of publications | Publications within the year 2022–2024. | Publication beyond the year of 2022–2024. |
| Search terms | Availability of specified search terms in either of abstract, title, or keywords of publication. | Non-availability of specified search terms in either of abstract, title, or keywords of publication. |
| Language of publication | Published in the English language. | Published in a non-English language. |
Data Analysis
Following data collection, the researchers attempted data screening. For this task, the investigators intended to co-create the research output by leveraging the AI writing tool Rayyan (https://www.rayyan.ai/) to get a preliminary overview of the SLR. However, the researchers could not benefit much from the AI writing tool in their attempt. One possible reason might be the initial search from the Scopus database, which employed the AI technique at the beginning of the research. Secondly, our dataset was comprised of only 23 documents, and there was hardly any scope for further data screening (such as PRISMA or other methods).
Recognizing the minute details of Rowe and Ngwenyama (2024) for data analysis, human resources, rather than AI, that take the center stage. Each document was employed for its content analysis to identify relevant themes. To locate the themes, emphasis was placed on specific sections of the document, including title, abstract, keywords, results, and conclusion. In this way, a total of 22 themes were identified from the sample studies. The contour of themes and the research dataset are appended to the article (Annexure 1–4). The themes are academic writing and publishing, academic integrity, plagiarism, cheating, academic dishonesty/academic misconduct, personalized learning, transformative learning environment, privacy and bias, policy framework, policy guidelines, legitimacy, ethics, copyright, detection, risk/threat, accessibility, equity, assessment, critical thinking, responsible use, pedagogy, and AI literacy. For a precise presentation of the results, the 22 individual but inter-twined themes were clustered into six broad themes (Table 2). Such clustering was necessitated to align each micro-theme with the specified RQs and accompanying broad-themes.
Table 2
Cluster of broad-themes and relevant micro-themes.
| RQ | Accompanying broad-themes | Relevant micro-themes |
| RQ1 | Viability of academic integrity | Academic writing and publishing, academic integrity, academic dishonesty/academic misconduct |
| RQ2 | Plagiarism and cheating | Plagiarism, cheating |
| RQ3 | Curriculum and assessment | Pedagogy, assessment, detection |
| RQ4 | Individual, institutional, and technical concerns | Personalized learning, critical thinking, interactive learning environments, accessibility, equity, privacy and bias, risk/threat |
| RQ5 | Ethical and legitimacy concerns | Ethics, copyright, responsible use, legitimacy |
| RQ6 | GAI policy matters | Policy frameworks, policy guidelines, AI literacy |
The process of performing thematic analysis in qualitative research is not standardized (Joffe, 2011). “Furthermore, in thematic analysis importance of a theme is not necessarily dependent on quantifiable measures, but rather on whether it captures something important in relation to the overall research question” (Vaismoradi et al., 2013, p. 402). The sample documents in this study had explicit references to the themes. The RQs’ in-depth examination of the research challenge was the main focus. Thus, for data analysis in this study, researchers rely more on the themes’ contour plot (like Milatz et al., 2015) than on the coding and reliability measures (unlike Hemmler & Ifenthaler, 2024).
Results
RQ1. What is the viability of academic integrity in higher education at the age of GAI?
In the context of GAI in higher education, many articles have expressed their concerns related to academic integrity and its viability. Our study has addressed the issue of positioning academic writing and publishing as the centrality of the entire discussion. Bozkurt (2024) investigates the complex interplay between GAI and human intellect in academic writing and publishing. The author examines the ‘organic versus synthetic’ paradox, emphasizing the implications of using GAI tools in educational and academic integrity contexts. The paper advocates for transparent reporting to uphold academic integrity. De Silva et al. (2023) perceived a primary challenge concerning the impact of ChatGPT on the academic integrity of scholarly work, where AI-generated content can be useful and detrimental in both teaching and research. The ability of the GAI tools to generate human-like responses with minimum instructions generates a grave concern among educators about students’ use of these tools for academic writing, which may constitute a breach of academic integrity (Kong et al., 2024).
A pool of researchers has identified the challenge of academic integrity with the usage of GAI on numerous occasions, such as Australian university educators’ perspective (Lee et al., 2024), business education (Fegade et al., 2023), challenges of GAI in education (Wang et al., 2024), the challenge of GAI for higher education as explained by ChatGPT (Michel-Villarreal et al., 2023), Co-regulated learning (Lodge et al., 2023), digital watermarking (Lancaster, 2023), higher education crisis (Song, 2024), nursing education (Simms, 2024), pedagogical model designing (Kong et al., 2024), promoting assignment integrity (Shanto et al., 2023), Turing’s imitation game (Matthews & Volpe, 2023). Sullivan et al. (2023) posited their research on Australia, New Zealand, the USA, and the UK to find out the effect of ChatGPT in higher education. One of the key themes that emerged was the issue of academic integrity. Per Watts et al. (2023), the emergence of GAI technology raises concerns regarding the value of open-ended writing tasks in the classroom. One such concern involves academic integrity and whether students will use these chatbots to produce sufficient responses to open-ended writing tasks. Locating ChatGPT, De Silva et al. (2023) perceived a primary challenge with its impact on the academic integrity of scholarly work, where AI-generated content can be useful and detrimental in both teaching and research. In a progressive outlook, Bozkurt (2024) discusses GAI’s diverse roles and advocates for transparent reporting to uphold academic integrity. While dealing with the problem, Bukar et al. (2024) precisely argued that focusing on specific capabilities of ChatGPT, such as developing tools to detect plagiarism and misinformation, may enhance resilience to such issues.
Any form of breach of academic integrity leads to academic dishonesty or academic misconduct. Through a multicultural study, a strong correlation was established between the cultural dimensions and the respondent’s view of GAI potential as academic dishonesty (Yusuf et al., 2024). The education community expressed its high concern about academic integrity, in that situation, Hassoulas et al. (2023) tried to investigate the reasons that constitute academic misconduct while dealing with such emerging technologies. With a perception that students will cheat in assignment submission with ChatGPT usage, a study was commissioned to seek ChatGPT’s advice on how to cheat in university exams. They further examined the workability of ChatGPT’s suggestions (Spennemann et al., 2024). To combat the higher education crisis and academic misconduct with GAI researchers argued for developing a clear institutional guideline (Shanto et al., 2023; Song, 2024). Even after accepting the concern of academic dishonesty posed by such large language models (LLMs), Blake (2024) advocates for the constructivist integration of GAI in education. Similarly, Lim et al. (2023) suggest embracing innovation with institutional governance and strategic regulation. The way educational bodies and institutions conceptualize academic misconduct may need to be rethought as well, they opined.
RQ2. What could be the impact of GAI on plagiarism and cheating in higher education?
Plagiarism and cheating continue to be the top worries for the higher education community in the wake of GAI. In addition, according to Bukar et al. (2024), giving students access to ChatGPT puts them at risk for plagiarism and cheating. Song (2024) makes it apparent that an elevation of these problems is caused by college students’ use of GAI tools. During an interview with ChatGPT, plagiarism concerns surfaced as one potential explanation for university bans on ChatGPT use (Michel-Villarreal et al., 2023). Shanto et al. (2023) propose a framework to prevent plagiarism, which will be covered in more detail later.
The rise of GAI has raised valid questions about academic integrity and, in particular, how easily these technologies could result in assessment cheating (Lodge et al., 2023). Spennemann et al. (2024) conclude that ChatGPT can be used successfully as a brainstorming tool to offer cheating advice through an inverted moral discussion. However, the effectiveness of this tool relies on the assignment markers’ attentiveness and the cheating student’s capacity to discern between options that seem feasible and those that are not.
RQ3. How far can GAI influence existing curriculum and assessment practices to inhibit academic dishonesty in higher education?
Numerous researchers have called for changes to the current curriculum and assessment procedures in light of the accessibility of GAI tools and the sustainability of academic integrity. A key element of curriculum transactions is pedagogy. To emphasize the necessity of updated instructional approaches in the AI era, Lodge et al. (2023) explore the link between students and AI. The researchers advocate for a more comprehensive strategy that places a strong emphasis on learning co-regulation and SRL. The idea is that teachers may better prepare students for an interconnected, AI-driven world by encouraging SRL and comprehending co-regulated dynamics. Blake (2024) argues against a regressive position and supports the positive inclusion of AI into educational processes, despite worries about academic dishonesty made possible by LLMs. Kong et al. (2024) use text-based GAI tools to create a pedagogical paradigm for SRL in academic writing. The model is intended to direct the creation of educational interventions and assist instructors and students in embracing GAI in the teaching-learning process. Simms (2024) offers a two-pronged strategy that uses AI as a scaffold for critical thinking development while integrating constructivist theories with Vygotsky’s Zone of Proximal Development framework.
Another crucial component of instructional design is assessment. There is room for creative assessment design when it comes to ChatGPT’s use in higher education (Sullivan et al., 2023). The goal of Hassoulas et al. (2023) is to examine marker accuracy in distinguishing between student-written university scripts and those generated with ChatGPT. Based on their findings, the researchers suggest responsible ways to incorporate GAI into assessment methods. Nearly half of respondents to a survey conducted at a top Australian institution use AI in their teaching responsibilities, with assessment design changes being the most common design change (Lee et al., 2024). Academics require assistance in constructing tests that successfully provide solid proof of students’ mastery of learning objectives because they are unsure of the existing capabilities of GAI. Institutions must, therefore, evaluate the suitability of their assessment designs as well as other related issues (Matthews & Volpe, 2023). Simms’s (2024) dual-purpose method consists of hands-on exercises that allow students to critically interact with AI-generated content, strengthening clinical judgment and preparing them for contexts in healthcare where AI is commonplace. Citing an organic chemistry assignment, research shows that chatbots react differently to various tasks. Additionally, chatbots generally do not engage in mechanical thinking at the same level as students (Watts et al., 2023).
Institutional procedures frequently use plagiarism detection during assignment submissions to assess whether there has been a violation of academic integrity. Bozkurt (2024) emphasizes the need for new methods to detect the use of GAI and urges a more thorough investigation of GAI tools. Focusing on particular ChatGPT features, including creating tools to identify plagiarism and false information, may improve resilience in some areas (like academic integrity), claim Bukar et al. (2024). An introduction to text-generation tools is given by Lancaster (2023), who also discusses the kinds of assessments that students can outsource, demonstrates the types of prompts that can be used to generate text, and shows one potential watermarking technique that might make it possible to identify generated text. Using an inverted moral valence, Spennemann et al. (2024) demonstrate how effective the suggested solutions are and how much GAI may be utilized to get advice on how to avoid discovery while commissioning and submitting contract-written assignments.
RQ4. What could be the individual, institutional, and technical concerns in accommodating GAI in higher education?
The process of making accommodations for GAI in higher education could have a variety of effects on the state of education today. Individual (critical thinking, individualized learning), institutional, or technical factors (accessibility, equity, privacy and bias, risk/threat) may be affected.
Blake (2024) emphasizes education in his description of GAI’s effects on the teaching-learning system, demonstrating how it may be used to customize content and change learning settings. Kong et al. (2024) demonstrate a self-directed learning design in academic writing. The significance of self-regulated and co-regulated learning is equivocated by Lodge et al. (2023). Personalization and interaction are highlighted by Wang et al. (2024) while talking about the technological affordances of GAI in education.
A pool of research has documented the use of GAI tools and the growth of critical thinking abilities. It is clear that students are developing their critical thinking abilities (Kong et al., 2024; Simms, 2024). The literature also contains a variety of report types. Evidence regarding the effectiveness of the increased critical thinking strategy is available (Kong et al., 2024). Students’ needs for critical analysis in some subjects may not be met by an over-reliance on ChatGPT (Michel-Villarreal et al., 2023).
Bozkurt (2024), focusing on institutional reforms, emphasizes the urgent necessity to use creative AI-inclusive techniques to adapt and change academic processes. A renowned Australian university serves as a lens for examining the effects of GAI in higher education. More than three-quarters of employees said they would like assistance, and less than a quarter believed the institution had prepared them sufficiently for AI (Lee et al., 2024). Management educators’ paradoxical viewpoint (Lim et al., 2023) provides a vivid theoretical description of the transformative learning environment. Other scholars have studied the topic to varying degrees (Blake, 2024; Lodge et al., 2023; Wang et al., 2024).
Accessibility is one of the technical issues that is emerging in this setting. When it comes to GAI’s accessibility and enforced limitations, management educators see a contradiction (Lim et al., 2023). According to Song (2024), AI literacy is seen as an essential instrument for guaranteeing equitable access to technology and avoiding the digital divide. Accessibility and the growing digital divide continue to be issues with technology affordances and the difficulties of GAI in education (Wang et al., 2024). According to Bukar et al. (2024), the use of GAI tools in higher education may increase the risk of educational inequality and the digital divide. Concerns about bias and data privacy have been raised by the training and operation of GAI tools (Fegade et al., 2023; Wang et al., 2024). Academic integrity is at risk due to the use of ChatGPT, according to a group of researchers (Lancaster, 2023; Sullivan et al., 2023; Zhong et al., 2024). Per Zhong et al. (2024), there are other risk variables linked to excessive reliance, teaching position, information transmission and knowledge level, ethical awareness, and ethical dangers.
RQ5. What are the ethical and legitimate concerns coming into the picture while attempting to locate GAI in higher education?
The section identifies difficulties related to the legitimacy, copyright, ethics, and appropriate use of GAI in higher education. The ethical and legal ramifications of authorship and ownership are discussed by Bozkurt (2024), particularly in light of copyright regulations and AI-generated work. According to Bukar et al. (2024), laws about ChatGPT have second-order impacts that might have both beneficial and bad repercussions. Fegade et al. (2023) looked for GAI use rules that were morally and responsibly conducted. Academic integrity has been a legitimate problem since GAI’s development (Lodge et al., 2023). These authors advocate for the ethical use of GAI in higher education: Matthews and Volpe (2023); Shanto et al. (2023); Song (2024); Wang et al. (2024); and Yusuf et al. (2024).
Given the complexities, Hassoulas et al. (2023) advise thinking about how GAI might be ethically incorporated into evaluation techniques and broadening our understanding of academic misconduct to take into account this emerging technology. According to Michel-Villarreal et al. (2023), there must be explicit rules for using ChatGPT responsibly in higher education. Yusuf et al. (2024), based on the multicultural perspective, contended that while the ethical use of AI and the prevention of academic dishonesty may necessitate strong policies that take cultural norms into account, the responsible use of GAI tools might improve learning processes.
RQ6. What does the existing body of literature convey on the GAI policy matters towards tackling the challenge of academic dishonesty in higher education?
The availability of writing resources and upholding academic integrity continue to be the top concerns for scholars and practitioners in higher education following GAI. Clear policies, guidelines, and a framework are, therefore, desperately needed (Michel-Villarreal et al., 2023; Song, 2024).
Two relevant primary themes emerged from the analysis of the research dataset’s content: the GAI framework and the GAI guidelines. Furthermore, there are distinct studies in which the need for AI literacy is mentioned. Researchers who adopt a holistic approach choose to take into account each of the three themes for the current conversation.
The study of English medium instruction (EMI) in general and GAI in particular is still ongoing. A GAI academic integrity policy development framework for global higher education cultures is presented by Bannister et al. (2023). The mixed method research provides a blueprint with four dimensions and nineteen descriptors to address the unique policy/practical features of GAI for EMI in tertiary education, using a virtual nominal group strategy. The most widely used GAI writing tool is still ChatGPT. A GAI policy-making approach based on the risk, reward, and resilience framework is presented by Bukar et al. in 2024. To provide components for the suggested decision-making framework, important aspects of ChatGPT concerns and motivations are methodically categorized under the risk, reward, and resilience categories. Using text-based GAI tools, a pedagogical design that is centered on the authoring cycle and SRL fosters students’ critical thinking and self-regulation while they compose academic writing. There are six interactive stages in the design, informed by Kong et al. (2024). Shanto et al. (2023) provide another conceptual framework, called PAIGE, that highlights the ethical integration of GAI, encourages active student participation, and fosters chances for peer learning experiences to preserve assignment integrity.
There is a substantial body of research that discusses GAI guidelines for enhancing academic integrity and instructional activities in HEIs. These recommendations are occasionally given as potential lines of inquiry for further study. De Silva et al. (2023), Lancaster (2023), Lim et al. (2023), Sullivan et al. (2023), and Wang et al. (2024) are among the researchers in the pool. The presentation’s prodigality restricts our ability to fully clarify the guidelines.
The crisis of academic dishonesty is expected to be lessened if AI literacy improves. Fegade et al. (2023) and Song (2024) placed a greater focus on AI literacy among students and/or teachers. However, the agenda is unable to produce a unified consensus. The technical aspects of learning, for example, are viewed with skepticism by Lodge et al. (2023), who instead place greater trust in the human-engaging aspects of learning.
Discussion
The viability of academic integrity about GAI adoption in higher education is identified in RQ1. Academic writing and publication, academic integrity and its potential violations, academic dishonesty, and academic misconduct are all linked to academic viability. Balalle and Pannilage (2025) examine the dual role of AI tools in promoting and compromising academic integrity, which is comparable to our critical stance on the current study problem. In contrast to our reports on the many forms of academic integrity viability, Sozon et al. (2024a) concentrate on identifying contributing reasons for academic integrity violations and suggested actions to alleviate the situation in higher education. Academic misconduct and academic integrity have been correlated. Thematic evolution shows that the focus has shifted from academic misconduct-integrity to problems like online fraud (Suprapto et al., 2024). In such a context, Surahman and Wang (2022) emphasize the importance of trustworthy assessment in order to address the problem of academic dishonesty in online education. According to Pilaez-Sanchez et al. (2024), the issue seems to be extremely concerning for school education, higher education in general (Liu et al., 2023), and other professional education fields (e.g. Miao et al., 2024).
As RQ2 provides, plagiarism and cheating with GAI assistance are serious concerns to academic integrity in action. Partly aligning with our observation, a report on the causes of plagiarism and cheating in higher education as well as solutions is included by Sozon et al. (2024b). Ahsan et al. (2022) review the literature on contract cheating and offer a conceptual framework to address the issue in higher education. According to Miranda-Rodríguez et al. (2024) “interventions with practical elements, such as plagiarism detection, paraphrasing, citation skills, in addition to using software to identify similarities, may reduce plagiarism” (p. 1). Hutson (2024), however, urges a reconsideration of plagiarism in the age of GAI due to the evolving nature of academic writing, plagiarism, and intellectual property rights. The current findings are consistent with earlier research that was based on either pre-ChatGPT databases (Moya et al., 2023) or post-ChatGPT databases (Peláez-Sanchez et al., 2024).
According to the RQ3, GAI has significant room to change the current curriculum and evaluation practices to prevent academic dishonesty in higher education. An empirical study that looked at how GAI may improve social work education using curricular navigation and assessment design in undergraduate course assignments corroborated our findings (Reimer, 2024).
In-service teachers noted that the adaptability of AI tools not only facilitated the creation of compelling lessons but also significantly improved students’ listening and speaking skills. The findings advocate for the strategic incorporation of AI in EFL teaching, suggesting that such technologies can lead to innovative pedagogical approaches that are attuned to the diverse needs of learners. (Yeh, 2024, p.1)
According to research, AI tools help students customize their self-evaluation, enhance learning support, contribute to improved academic performance, and develop collaborative learning environments (Cai et al., 2024). According to Labadze et al. (2023), students benefit most from AI-powered chatbots in three main areas: help with homework and studies, individualized learning, and skill development. The primary benefits for teachers are enhanced pedagogy and time-saving support. Thereby, the findings call for a shift in higher education’s existing pedagogy (Simms, 2024) and evaluation procedures (Foung et al., 2024).
To accommodate GAI in higher education, RQ4 addresses individual, institutional, and technical challenges. Regarding individual concerns, the results of this study corroborate those of Liu et al. (2023) about the views of critical thinking in a different context and individualized learning (partially). The findings of Moya et al. (2023) are combined with the results of institutional services such as creating an interactive learning environment and providing parity in the use of GAI. Urdan and Marson (2024) are completely in line with the current research output on technological aspects of accessibility and risk/threat related to data privacy and bias. Sozon et al. (2024a) highlighted five factors—individual, institutional, societal, cultural, and technical—while discussing the causes of academic integrity violations in higher education. Therefore, our results partially align with the relevant report.
The RQ5 focuses on the legitimacy (and copyright) issues and ethical (responsible) use requirements of GAI in higher education. The status of AI and academic integrity in higher education is investigated in a rapid scoping review. The findings demonstrate both limited and unrestricted ethical implications of AI. AI for cheating, AI as legitimate support, an equality, diversity, and inclusion lens into AI, and new suggestions to address AI’s ramifications in higher education are among the views presented (Moya et al., 2023). To find out how students and faculty at Lebanon’s universities see the role of AI tools, a qualitative study was carried out. The results urge the creation of precise rules and regulations for their ethical and responsible usage (Karkoulian et al., 2024). In the context of higher education, separate research by Bozkurt (2024) and Pilaez-Sanchez et al. (2024) offers comparable reports.
The RQ6 captures the need for GAI policy regulation in higher education. To address the opportunities and problems it presents for higher education, researchers invited ChatGPT. It highlights the urgent necessity for precise policies, guidelines, and frameworks to properly integrate ChatGPT in higher education while offering solutions to overcome challenges (Michel-Villarreal et al., 2023). Nagpal (2024) made recommendations for universities, instructors, students, researchers, and staff to establish a comprehensive GAI guideline by comparing the policies and guidelines of the top twenty Canadian institutions as determined by the Times Higher Education World Rankings. In terms of the structuring and schematizing of the suggestion to degrees, the present study found some similarities and variances. According to a study on 40 institutions’ inclusive GAI plans from six different global areas, our policy advice was found to have mixed analytical results (Jin et al., 2025). Connecting to the conversation above, Rudolph et al. (2024) recommend AI literacy among teachers and students.
What More to Carry Forward?
RQ1–6 illustrate how various viewpoints on the use of GAI in higher education and academic dishonesty interact. The fact that individual RQ’s concerns are intertwined with those of others highlights the significance of entangled thing ethnography (Murray-Rust et al., 2019) in this study.
It is seen in RQ6 that the GAI framework’s output varies depending on the specific research topic (Bukar et al., 2024; Kong et al., 2024; Shanto et al., 2023). It may be argued that creating a new GAI framework may not remain equally appealing in all circumstances. Rather, there exists some overlap among the current GAI guidelines. In those investigations, the technique and/or the research goal were different. We suggest creating a comprehensive guideline (Table 3) to address academic dishonesty with GAI usage in higher education to support the existing limitation. We prefer to communicate the new guideline in three directions—individual, institutional, and technical concerns—by using an inductive approach (Lee et al., 2024). Since the new guideline is based on existing material, no additional discussion is included. Five elements that lead to academic integrity violations in higher education were emphasized by Sozon et al. (2024a) in the discussion of RQ4 results. It is believed that social and cultural elements are more under the influence of an academic institution’s external environment. The said norms, however, can also be imprinted in an institution to varied degrees. By using this logic, we limited our discussion to individual, institutional, and technical concerns.
Table 3
Comprehensive GAI guideline for GAI usage in higher education.
| INDIVIDUAL CONCERNS | INSTITUTIONAL CONCERNS | TECHNICAL CONCERNS |
|---|---|---|
| Personalized learning skills development | Specifying discipline-specific AI interventions | Enhanced accessibility, ensuring equity to AI access |
| Self-directed learning skills development | Change of existing curricula and assessment practices, development of a clear institutional GAI policy | Minimizing data biases, protecting the data privacy of the users |
| Critical thinking skills development | Student and staff training on GAI | Framing copyright protection law, use of AI tools for plagiarism detection |
| Student-teacher-AI integration in teaching and learning | Fostering AI inclusive learning environment | Clearing GAI legitimacy issues |
| Ethical/responsible use of AI | Promoting awareness of academic integrity | Spreading AI literacy among the stakeholders |
Educational Implications, Limitations, and Scope for Future Research
The current study offers a thorough overview of academic dishonesty in higher education throughout the GAI era. This study offers a more contemporary picture than Moya et al. (2023), who used the pre-ChatGPT timeframe to finalize their dataset. The new study can also serve as a supplement to Liu et al.’s (2023) contextualized study on school education.
Comprehensive guidelines for using GAI to address academic dishonesty in higher education are given in Table 3. We believe the guideline is more complete now after including three distinct concerns in its framing, and will remain instrumental in addressing the issue of academic dishonesty in higher education. According to Kong et al. (2024), the guideline could assist educational stakeholders in making the most of GAI tools’ potential in higher education practices.
However, the research is not perfect. The qualitative approach used for the research makes it vulnerable to individual bias of the researchers in the data analysis. Future studies could use different research methods to confirm the limitations and remove the potential for researcher bias (Hemmler & Ifenthaler, 2024).
Thematic analysis is the method used in the literature review. What might occur if different kinds of study methodologies were used (Sareen & Mandal, 2024)?
Research indicates that the quality of the study is more significant during a review than the total number of studies ultimately taken into consideration (Moya et al., 2023). However, if a larger dataset is utilized to review the subject, the current findings may alter. Future inquiry may be conducted to evaluate the proposed GAI guideline’s efficacy.
Conclusion
Addressing academic dishonesty in higher education during the GAI era is the aim of the study. The research is methodically guided by six RQs. The viability of academic integrity with GAI use in higher education is captured in RQ1. According to RQ2, the majority of these issues stem from GAI assistive writing tools and manifest as plagiarism or cheating. There is substantial room for improvement in the current curriculum and evaluation procedures to prevent academic dishonesty in higher education (RQ3 provides). RQ4 addresses GAI’s individual, institutional, and technical concerns to transparently accommodate it inside the higher education environment. RQ5 highlights additional considerations, such as issues of legitimacy and ethics. As demonstrated by RQ6, there is no one-size-fits-all GAI framework for addressing the issue of academic dishonesty. The GAI guidelines of the relevant database show some degree of alignment.
To make the most of GAI’s potential in higher education, a comprehensive guideline is included. The study calls for practical implications of the new GAI guideline in higher education to make the research output more engaging. For the stakeholders in higher education, educational consequences are appended. The effectiveness of the new GAI guideline may be examined in future studies using a different methodology (other than thematic analysis), approach (other than qualitative research), or both. The study’s conclusions offer a thorough understanding of academic dishonesty associated with GAI in higher education. The research is presented from a critical yet optimistic standpoint. The study’s outcome is to create a solid ethical foundation for the application of GAI in all academic endeavors.
Data Accessibility Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Sustainable Development Goals (SDGs)
This study is linked to the following SDG(s): Quality education (SDG 4), Industry, innovation and infrastructure (SDG 9), Reduced inequalities (SDG 10).
Additional File
The additional file for this article can be found as follows:
Supplementary Material
Contour of Themes and The Research Dataset: Annexure 1–4. DOI: https://doi.org/10.55982/openpraxis.17.2.820.s1
Ethics and Consent
The research was not done on any animal or human body, so no special ethical consideration was taken. However, conventional research ethics were followed when conducting this study.
Competing Interests
The authors have no competing interests to declare.
Author Contributions (CRediT)
Ramesh Chander Sharma: Conceptualization, methodology, formal analysis, investigation, data curation, visualization, writing—original draft preparation, writing—review and editing; Suman Kalyan Panja: methodology, formal analysis, supervision, project administration, writing—original draft preparation, writing—review and editing. All authors have read and agreed to the published version of the manuscript.