The AI-powered learning loop in higher education

Artificial intelligence (AI) continues to impact higher education in various aspects, including learning and teaching, assessment, grading, skills for future work and the future career of graduates. AI can achieve more efficiency than human beings, particularly in learning and teaching. It is likely to transform education by personalizing teaching and providing immediate feedback (Zouhaier, 2023). For Zhang (2023), the continuous advancement of AI leads to a diversification of teaching evaluation methods, the personalization of learning experiences and greater inclusivity within the education ecosystem. Similarly, Al-Zahrani and Alasmari (2024) assert that the future of higher education will be potentially characterized by the ethical integration of AI, collaboration and ongoing support for lifelong learning.

With the continuous advancement of AI applications, students may prioritize their interests and hobbies over training for careers that robots will ultimately take over (Zouhaier, 2023). While considering the ethical challenges of AI-based education, students leverage AI tools to acquire new skills that may advance their future careers (Hamdi, 2024). AI helps in reducing the learning disparities between students, thereby increasing their engagement and academic performance. Fawns et al. (2024) indicated that students are using generative AI to accomplish various complex tasks, e.g. writing, editing and demystification of complex notions, while acknowledging the “tensions” associated with the use of these tools, particularly their concerns about meeting the academic integrity expectations set by their institutions. Another tension expressed by students pertains to the uncertainty surrounding the extent to which the use of AI use is tolerated in their assessments. Therefore, Fawns et al. (2024) urge higher education institutions to adopt well-rounded and pragmatic academic policies that clarify the “acceptable use” of AI tools, balancing the gains of efficiency with academic integrity considerations. According to Saidakhror (2024), universities must align with the developments of AI through the enhancement of their infrastructure and policies. Asri (2024) contends that university curricula need to be aligned with ongoing updates of AI applications.

Drawing on these recommendations, the present research proposes to examine the relevance of an educational approach adopted at a business college in a private university in Kuwait that consists of adjusting the assessment policy to permit students’ use of AI tools in developing and editing homework assignments while allocating a significant evaluation component for the individual discussion of the submitted content.

To our knowledge, no prior research has assessed the tangible outcomes resulting from adopting a less coercive approach toward students using AI tools in developing their homework assignments in a Middle Eastern higher education context, i.e. Kuwait.

Despite the growing number of recent scholarly publications focusing on the role of AI in higher education, the theoretical perspective has been largely prevalent, thereby underscoring the necessity for evidence-based research (Marengo, Pagano, Pange, & Soomro, 2024). Additionally, student utilization of generative AI tools is anticipated to increase further in the coming years (Amoozadeh et al., 2024).

The significance of investigating the impact of AI-policy amendments at the chosen college stems from the complexity of monitoring students’ work integrity using specialized detection applications. College officials recognize that similar practices have become increasingly difficult with the emergence of AI tools that produce highly humanized outputs. Instead, the college aspiration – through the application of the updated policy – now focuses on fostering the spontaneous adoption of effective and ethical learning strategies among students. Previous studies have shown that students in the region often experience strong social pressure to reach exceptional results, which can influence their choice of learning approaches. Such pressure makes it essential to create educational environments that promote appropriate and constructive practices (Abulibdeh, Zaidan, & Abulibdeh, 2024; Aljurf, Kemp, & Williams, 2020; Alsuwaileh, Russ-Eft, & Alshurai, 2016; Kadayam Guruswami et al., 2023).

The objective of this study is to assess the utility of taking a practical approach in higher education institutions to AI-assisted assignments among students, which may stimulate their assimilation capacity and allow them to devote more time to understanding and mastering the information gathered due to efficiency gains. By offering a context-specific case, policymakers in other parts of the world might use the findings of this study to proactively change their educational philosophy and academic policies in light of the rapid growth of AI tools in general. Specifically, the outcomes of the present research may contribute with a deeper understanding that can inform other higher education institutions, regionally and globally, in catalyzing the integration of generative AI tools into holistic pedagogical approaches to achieve sustainable student learning outcomes.

Hence, the following research questions are posed to guide the explorative study that would result in research propositions underlying a comprehensive framework:

1. How does a tolerant approach to AI-assisted assignments in higher education impact business students’ assimilation capacity and learning efficiency?

2. What implications does a tolerant approach to AI-assisted assignments have for academic policy development in higher education?

The paper is organized as follows. A literature review is provided first, followed by an overview of the methodology applied to this study. Then, the results of the content analysis are presented and discussed. Lastly, the conclusion highlights the main theoretical and practical contributions alongside avenues for future research.

Educators across all generations and disciplines are open to incorporating AI tools into their educational practices. They acknowledge that such technologies can enhance teaching processes by facilitating personalized learning, automating assessment, providing chatbot assistance as well as virtual/augmented reality experiences (Annuš, 2024). AI can promote personalized education via the analysis of student data to generate customized learning plans. For instance, the use of chatbots and AI-assisted virtual assistants has been shown to improve learning outcomes in medical science education. Moreover, AI supports the democratization of education through different tools such as real-time translation and adaptive learning materials (Domanchuk & Chornenka, 2023). AI-powered systems also assist in automating various operational processes in higher education institutions, such as student admissions and registration, student services and course scheduling. This automation enhances organizational efficiency and enables the reallocation of resources to strategic functions that have a more significant impact on student learning outcomes (Stoyanova & Angelova, 2024).

Integrating AI into educational systems and pedagogical practices necessitates the implementation of appropriate safeguards to address several hazards regarding data confidentiality and privacy, system security, data bias and plagiarism. Certain AI applications require the collection of personal information from students. Furthermore, the data utilized to generate AI content is sourced from public sources, which may lead to the dissemination of biased information among users. Also, the increasing prevalence of AI tools in education can deter human connection in the educator–student relationship, thereby hindering the acquisition of socio-emotional skills among students and the sense of community within the educational group (Kamalov, Santandreu Calonge, & Gurrib, 2023). Despite the significant effect of generative AI tools on students’ divergent thinking, they are associated with adverse effects on creativity and creative confidence (Habib, Vogel, Anli, & Thorne, 2024). The relentless development of AI tools has sparked intense debates regarding the ethical considerations associated with their use, as well as concerns about the uncertain future of teaching and learning (Bahroun, Anane, Ahmed, & Zacca, 2023).

Different measures are proposed in the literature to deal with the challenges associated with integrating AI into education. These include the continuous redesign of curricula, the updating of educational methodologies and the benchmarking of successful practices within the sector (Abulibdeh et al., 2024). Higher education institutions need to undertake a systemic approach to fully harness the transformative potential of AI, hence alleviating complexities and potential pitfalls. However, the variations among educational institutions in terms of policies and student perceptions further complicate the development of a unified understanding about the appropriate use of AI (Xiao, Chen, & Bao, 2023). The strong resistance to AI adoption among academics and administrators in higher education institutions presents an additional challenge in this regard (Stoyanova & Angelova, 2024).

Students report that generative AI tools assist them in brainstorming activities. These tools have expanded the range of their ideas, generated in a more fluent manner with greater elaboration (Habib et al., 2024). generative AI tools exert a significant impact on student’s entrepreneurial competencies, such as the identification of business opportunities, the adoption of creative and visionary thinking patterns, the assessment of business ideas and sustainability awareness (Somia & Vecchiarini, 2024). Business students using generative AI chatbots can more easily retrieve market data as well as pertinent information on customer preferences and competitor positioning (Mavlutova, Lesinskis, Liogys, & Hermanis, 2020). However, students express concerns about the originality of the ideas produced and the potential risk of unintentionally violating intellectual property rights (Habib et al., 2024), along with fears pertaining to data security and misinformation (Gasaymeh, Beirat, & Abu Qbeita, 2024). Furthermore, students’ application of generative AI tools in the writing process can be detrimental to their writing skills (Puxon, Brook, & Prevatt-Goldstein, 2024).

Another critical issue emerging in the literature relates to academic integrity. Students frequently report uncertainty about where the boundaries lie between appropriate and inappropriate use of generative AI, which creates confusion and anxiety in their learning journey (Fawns et al., 2024). This ambiguity increases the risk of unintentional academic misconduct, particularly in contexts where institutional policies remain vague or inconsistently enforced (Xiao et al., 2023). Integrity concerns are heightened by the fact that many AI systems generate convincing but unverifiable sources, which students may unknowingly reproduce in their assignments.

Studies in the Middle East already highlight academic dishonesty as a persistent issue, often shaped by cultural, institutional and social pressures to achieve exceptional results (Aljurf et al., 2020; Alsuwaileh et al., 2016). The introduction of AI tools further complicates this picture by making it more difficult to distinguish between authentic student work and AI-assisted outputs. Kadayam Guruswami et al. (2023) note that health-professions students in the region expressed serious concerns about blurred lines between acceptable use and misconduct, signaling that the integrity challenge is not limited to a single discipline.

At the same time, some scholars emphasize that framing AI use solely as a threat to integrity risks overlooking its potential to cultivate more responsible and transparent academic practices. Amato and Schoettle (2023) argue that higher education institutions should develop clear, pragmatic guidelines that balance efficiency and innovation with ethical expectations, ensuring that AI is used to support learning rather than replace it. Similarly, policy-focused research stresses that institutions need to communicate explicit definitions of acceptable use, incorporate integrity training into curricula and align detection tools with the evolving nature of AI technologies.

Given the exploratory character of the study topic, a qualitative approach is used, based on group interviews with final-year students enrolled in a Bachelor of Management program at a college of business in Kuwait who were observed in multiple in-class activities completed with the permitted use of generative AI tools. In the following sub-sections, additional information is provided regarding the context of the study, including the subject institution and its AI policy, participant profiles, data collection and data analysis.

The College of Business was established in 2004 as part of one of the first private universities in Kuwait. The university offers 24 diploma and bachelor’s degree programs across different disciplines, including business, engineering and aviation. As per the latest data, the university has a total capacity of 2,600 students, with 430 enrolled in business programs. Business specializations include the following: management, marketing, human resource management, accounting, management information systems, events management and banking and finance. The vast majority of students are Kuwaiti nationals of both genders, representing more than 95% of the total enrollment. International students are only those who already reside in Kuwait, primarily the descendants of expatriate workers. Indeed, students from abroad cannot register for any university program unless they have valid residency in Kuwait.

The College of Business employs an experiential learning approach based on project-based learning and is accredited by Kuwait’s Private Universities Council as well as the Accreditation Council for Business Schools and Programs (ACBSP). The teaching staff is internationally diverse, with nearly 70% being expatriates.

The College of Business has taken a more indulgent and progressive stance toward students’ use of AI tools in the preparation of homework assignments beginning in Fall 2024–2025. According to the college’s updated assessment policy, students will not be penalized if they can adequately explain the information offered in their reports during individual discussion sessions. Students are also expected to use verified material and cite valid academic and professional sources. The individual discussion component represents 30% of the assessment evaluation. This weight will be raised to 50% in the subsequent academic year 2025–2026. This policy reflects the college’s awareness of the growing role of generative AI tools in academic and professional contexts. Hence, the policy aims to promote responsible and ethical learning practices among business students.

Twenty students were observed and interviewed for the purpose of this research. The students were enrolled in a final semester mandatory course, which falls under the bachelor’s degree in management, namely “Contemporary issues in HR.” The profile of the participants in terms of gender, nationality composition and age groups is detailed in Table 1 below:

As indicated in the table above, all participants belong to the same age group and are evenly distributed across both genders. As noted in the college description above, the study participants were predominantly Kuwaiti nationals. Students from abroad are not allowed to join any public or private university program in Kuwait unless they hold a valid residency in the country.

The students were all enrolled in a course that addressed contemporary issues in human resource management. The course requires students to apply analytical thinking to evolving workplace trends while using textbook concepts and theories. During Spring 2024–2025, 20 students participated in five weekly in-class activities. These activities were completed and evaluated individually, totalizing a participation weight of 10% incorporated into the overall grade. All activities were created by the course instructor – the lead researcher and distributed in class after students completed specific reading assignments from the textbook. The activity questions required students to refer to and apply particular notions from the textbook. The purpose is to discourage students from relying entirely on the information provided by generative AI tools and to foster effective learning while ensuring relatedness to the lesson topics.

The research design applied the following three stages for each in-class activity:

1. Preliminary specification of the personal writing process and expected learning outcomes;

2. Conduction of regular in-class activities, allowing the use of generative AI, yet reference to the textbook information is necessitated; and

3. Semi-structured group interviews with students were conducted to explore their perceptions regarding the learning experience with a focus on their pains and gains in the writing process.

The group interviews with all participants included open-ended questions after each exercise. The lead researcher took notes on interview responses along with in class personal observations and confirmed them with the students to ensure content validity. The interview protocol did not require all students to answer every question. Instead, they were free to participate voluntarily and verbally in response to the questions of their choice. Answers were not recorded to allow participants to express themselves more freely and to alleviate concerns that recorded responses would be evidence of academic dishonesty. Personal observations focused on gathering insights about the study’s context, including the depth of topic comprehension, changes in student engagement, especially among less participative students, the level of independent learning as demonstrated by the number of questions exchanged with peers or the instructor and the frequency of using generative AI tools on mobile devices such as tablets or laptops.

The exercise instructions forced students to refer to textbook content, necessitating iterative connection with generative AI technologies to refine their answers. Each activity was followed by a discussion session engaging all students to assess the level of comprehension and provide clarity whenever necessary. To encourage critical thinking and discourage the easy replication of AI-generated solutions, each student was required to read and explain the written answer. The final results were shared with the participants to ensure the study’s reliability and to potentially collect further impressions and feedback. There were no objections to the final research manuscript.

The gathered responses and observations were examined after being encoded to identify common themes that will serve as the basis for an integrative conceptual framework. Following the recommendations of Creswell (2003) for qualitative studies, our interview data were transcribed and organized by source. Further, extracted data were screened to derive meanings and identify general themes, coded with labels. To align the study findings with existing streams of literature, the analysis of the data aimed at identifying relevant parallels with established theoretical concepts.

The experience led to various insights into students’ perspectives on AI utility in completing assignments. Firstly, the advantages pertaining to time efficiency were frequently emphasized in the responses collected from our participants. Students reported that the use of generative AI tools enabled the obtainment of more focused results in a reduced amount of time. As per the statements compiled in Table 2 below, AI tools tend to simplify numerous research tasks and channel students’ energy toward developing a stronger understanding of the topic (i.e. time efficiency and more focused information search). Table 3 identifies the key issues that arise while using generative AI in the given context.

More specifically, as shown in Table 2, gaining new insights was the most frequently cited benefit (11 students), followed by support for understanding complex information (9), while time efficiency and focused information search (6) and structuring thoughts (4) were also commonly mentioned. In contrast, Table 3 indicates that inaccuracies in AI-generated outputs were the most prevalent concern (8 students), whereas over-reliance on AI tools (4) and ethical concerns (3) were reported less frequently. These patterns suggest that while students recognized a broad range of benefits, issues of accuracy and trust remained their dominant challenges.

In this regard, generative AI technologies are anticipated to assist students in organizing massive amounts of information and deriving more structured thoughts to aid in their analyses (i.e. structuring thoughts). Generative AI may potentially reveal previously unseen elements of the issue under study. AI may provide rapid feedback on student responses before they are submitted. Respondents emphasized that AI can be an effective practice tool before exams.

Nonetheless, our interview findings show that delivering these benefits is heavily dependent on the users’ intentions, especially when the goal is to learn enrichment. Reliance on AI tools, on the other hand, may impair users’ cognitive skills if they rely excessively on the material without making a significant effort to integrate their own input. In this context, participants noted that copying AI content without rigorous review carries considerable hazards, such as inaccurate referencing and erroneous affirmations. As a result, it was stated that effective AI use is only possible when users have a genuine drive to learn and a strong sense of academic integrity (Table 3).

The results of the initial interview coding indicate a possible link between the use of generative AI tools and the students’ cognitive loads. Generative AI techniques save students time and effort when performing academic assignments, allowing for more introspective engagement that ultimately results in deeper learning. According to Amato and Schoettle (2023), AI has a positive impact on students’ cognitive abilities, prompting educators to emphasize its usage in improving individual learning over its restriction. Cognitive load theory was proposed thirty years ago (Sweller, 1988) and it is extensively used as a framework to develop instructional resources that correspond to students’ memory capacity (Sohrabi et al., 2023). Originating from psychology, the concept of cognitive load refers to the mental effort engaged by individuals in active and limited memory (Seeber, 2011). Three components of cognitive load are discussed in the literature: (1) intrinsic cognitive load, which refers to the mental effort necessary to comprehend the information itself. This load is impacted by the intricacy of the material and the learner’s prior understanding of the subject; (2) extraneous cognitive load, which refers to the mental effort required to grasp the layout of the information presented. This type of load is related to the instructional approach in place and the details offered and (3) germane cognitive load, which is necessary to optimize the assimilation of intrinsic load into the current mental models (Duran, Zavgorodniaia, & Sorva, 2022; Hochstrasser & Stoddard, 2022).

Several multimedia teaching materials have the disadvantage of overtaxing learners’ information processing capacities, reducing their active memory capacity (Salehi, Moradimokhles, Ghasemtabar, & Qarabaghi, 2017). As a result, teachers should seek alternate approaches to overcoming student memory constraints (Duran et al., 2022). In this regard, our study’s findings indicate a negative relationship between AI use and student cognitive load. The influence of AI tools on teaching and learning must be reviewed on a frequent basis to stay up with the rapid progress of these technologies (Philbin, 2023). A substantial amount of evidence supports the idea that AI reduces cognitive load by simplifying complex tasks, making knowledge more accessible and automating regular actions. For example, evidence suggests that AI-based data visualizations increase consumers’ cognitive load, which negatively correlates with their confidence in AI systems (Hudon, Demazure, Karran, Léger, & Sénécal, 2021). Explainable AI has a strong influence on physicians’ cognitive load, task performance and task duration (Herm, 2023). In the same vein, AI-enriched textbooks enhance deep learning among students by optimizing their cognitive load, particularly the germane component (Koć-Januchta et al., 2022). Furthermore, AI-assisted language learning strategies significantly reduce learners’ cognitive load and increase language acquisition efficacy. These strategies address specific learning dimensions such as the delivery of tailored feedback, the design of engaging exercises featuring speech recognition and the use of data-driven insights to generate intelligent tutorials (Feng, 2025).

Overall, we assume that adopting AI-tolerant assessment policies will reduce students’ cognitive load, leading to higher self-efficacy.

According to the lead researcher’s observations during the in-class activities, students’ quality of participation improved significantly when generative AI tools were permitted. Notably, they were more inclined to engage in debates and elaborate on their ideas with greater confidence. This capacity continued to develop as further exercises were conducted. A theoretical concept that best describes the increasing confidence in one’s skills and knowledge is self-efficacy, defined as an individual’s belief in their ability to perform specific tasks or behaviors successfully (Pekmezi, Jennings, & Marcus, 2009).

Higher self-efficacy is associated with positive outcomes, including better academic performance (Maddux & Kleiman, 2016). A recent study demonstrated that the development of AI capabilities among higher education institutions increases the levels of students’ self-efficacy and creativity. These AI capabilities revolve around the possession of resources, skills and consciousness about innovation and reform necessity (Wang, Sun, & Chen, 2023). In this sense, universities investing in AI-driven infrastructures and fostering a culture that values responsible and independent learning will be able to stimulate students’ confidence in their capacity to tackle complex tasks and formulate relevant solutions. Under such conditions, students will be able to achieve the intended learning outcomes and approach academic challenges with greater autonomy and confidence, which may result in better course evaluations.

The results suggest that the application of flexible AI policies has a favorable impact on tacit knowledge acquisition. Students claimed that AI helped them structure and formulate their ideas more effectively. AI tools provide instant feedback and structured guidance, with access to insightful examples. However, the internalization of these skills requires students’ active engagement and sustained critical thinking. Although further empirical research is needed to determine the relationship between AI and tacit knowledge acquisition, recent studies have revealed that ChatGPT helps second language learners improve their academic writing skills (Jacob, Tate, & Warschauer, 2023). AI-assisted feedback helps students enhance their writing skills by encouraging reflective thinking and iterative learning (Kim & Tan, 2023). Nonetheless, a high AI use frequency significantly increases knowledge acquisition gains for students (Harisankar, Malik, & Nagarkar, 2024). Together, these emerging patterns form the basis of what we refer to as the AI-powered learning loop, a conceptual model that links cognitive load, self-efficacy, tacit knowledge acquisition and academic performance.

Additionally, we argue that AI-aided tacit knowledge acquisition is likely to boost students’ self-efficacies. The relationship between the development of tacit knowledge and student self-efficacy has received little attention in research. In contrast, some studies have investigated the impact of self-efficacy on student learning outcomes (Meng & Zhang, 2023; Zheng, Chang, Lin, & Zhang, 2021). Generative AI tools can provide individualized and interactive learning experiences that incorporate quick feedback mechanisms. All of these features may encourage experimentation and the development of various abilities (e.g. conciseness, eloquence, problem-solving, presentation design), reinforcing students’ sense of competence, which is linked to their self-efficacy. As a result, investigating the relationship between students’ tacit knowledge acquisition and self-efficacy in the context of emerging generative AI tools could be a promising area for future research.

These five observations are summarized in the conceptual model presented in Figure 1, which illustrates the AI-powered learning loop linking cognitive load, self-efficacy, tacit knowledge acquisition and academic performance. This AI-powered learning loop offers a structured way to interpret how tolerant AI policies may reinforce sustainable student learning.

Our findings also highlight that academic integrity concerns remain deeply intertwined with the integration of generative AI tools in higher education. Although students in our study emphasized learning benefits, such as efficiency, deeper understanding and improved self-efficacy, several participants expressed uncertainty about the boundaries of acceptable AI use. This observation is consistent with prior research pointing to ambiguities in institutional policies and student perceptions (Fawns et al., 2024; Xiao et al., 2023). By adopting a tolerant but structured policy, one that requires students to explain and defend their AI-assisted work, the studied college offers a pragmatic model that aligns with recent calls for balancing innovation with integrity (Amato & Schoettle, 2023). Such an approach suggests the possibility that AI can be incorporated responsibly into pedagogical practices without undermining ethical standards, thereby reinforcing the sustainability of the AI-powered learning loop. Beyond these five observations, it is also important to consider the implications of a tolerant approach to AI use for management education specifically and higher education more broadly.

This study set out to investigate the impact of generative AI tools on business students’ academic performance, with a focus on their assimilation capacity, cognitive load, tacit knowledge acquisition and self-efficacy. This study employed a qualitative design, combining classroom observations and group interviews with 20 final-year business students in Kuwait. The analysis revealed that students perceived generative AI tools as beneficial for saving time, gaining new insights, structuring ideas and supporting the understanding of complex information, while also raising concerns about inaccuracies, over-reliance and integrity.

The proposed research model integrates the mentioned dimensions to explain how a tolerant approach to AI use in higher education can shape learning outcomes and processes. Termed the AI-powered learning loop, this model is underpinned by the existence of a robust learning orientation among students linked to frequent use of AI tools. These processes mutually sustain and enhance the learning loop. Thus, empirical validation of this model should consider variations in students’ learning orientation.

In addition to these theoretical contributions, our findings also carry important implications for management education and higher education policy. For management colleges, a tolerant approach to AI use, combined with assessment mechanisms that require students to explain and defend their work, can foster critical thinking, accountability and deeper learning rather than encouraging rote reliance on technology. Integrating AI into assignments, while emphasizing oral defense, reflective discussion and the use of verified sources, may both mitigate risks of academic dishonesty and enhance students’ professional readiness. Such practices align with the broader goals of business education, which emphasize adaptability, problem-solving and responsible innovation.

At the policy level, the study points to broader implications for higher education institutions. As the literature emphasizes (Fawns et al., 2024; Amato & Schoettle, 2023), the absence of clear guidelines can generate confusion and anxiety among students, whereas pragmatic and transparent AI policies can provide clarity and promote ethical use. Tolerant approaches that encourage responsible engagement with AI may help cultivate student self-efficacy, digital literacy and tacit knowledge acquisition. Institutions will also need to provide training for faculty, revise detection mechanisms and adapt policies to regional and cultural contexts. These implications directly address our second research question, underscoring how tolerant AI policies can inform institutional decision-making in both management education and higher education more broadly. The AI-powered learning loop thus provides both a conceptual and practical foundation for designing tolerant AI policies in management and higher education.

Although the proposed AI-powered learning loop cannot be generalized to other academic contexts, its exploratory nature paves the way for empirical work within the boundaries of the cognitive load theory. We believe that our conclusions – in their current scope – should not be tested in STEAM disciplines. Instead, they should primarily focus on the social sciences, particularly business studies. An interdisciplinary research initiative will facilitate the incorporation of diverse perspectives that will allow further model refinement, hence its potential extrapolation to scientific specializations. Overall, by examining the effects of generative AI tools on assimilation capacity, cognitive load, tacit knowledge acquisition and self-efficacy, this study underscores how tolerant AI policies can enhance student learning outcomes and inform the future of management education and higher education policy.

This study received ethical clearance from the Research Center of the College of Business, Australian University – Kuwait (Approval Date: March 4, 2025). Participation by students was voluntary and no personal data were collected. Group interviews were conducted without recordings and anonymized notes were used to protect confidentiality. The research was carried out in accordance with the ethical standards of the institution and the principles of voluntary informed consent.