Amid rapid technological transformations, artificial intelligence (AI) has emerged as a key branch of computer science, capable of simulating human cognitive functions such as learning, decision-making, and problem-solving. Generative AI, particularly large language models like ChatGPT developed by OpenAI, represents a transformative advancement by generating novel content—including text, images, and code—beyond traditional data analysis. These tools have become widely accessible, notably within higher education, contributing to digital transformations in teaching, learning, and research. Other platforms, such as Google Gemini, Grammarly AI, Jasper, and Perplexity, similarly offer advanced generative and creative capabilities, supporting academic writing, knowledge acquisition, and analytical synthesis (Gray et al., 2025; Gammoh, 2024; Altikriti and Nemrawi, 2025).

Artificial intelligence (AI) has reshaped critical aspects of contemporary life, notably in higher education, where it drives innovation in administrative, research, and pedagogical practices. The emergence of generative AI tools, such as ChatGPT and other OpenAI applications, has enhanced learning experiences, expanded opportunities for specialized education, and supported teaching and administrative processes. Universities increasingly utilize intelligent teaching applications and virtual assistants for predictive analytics, automated assessment, and personalized learning, while also employing AI in academic research through data analysis, content generation, literature reviews, and scholarly writing. Studies across universities in Jordan, Saudi Arabia, and the UAE indicate that around 88% of students use AI-based virtual assistants for assignments, examinations, language learning, and programming (Aldossary et al., 2024; Almulla and Ibrahim Ali, 2024; Alnsour et al., 2025; Al Kaabi, 2025; Al Mashagbeh et al., 2025; Vieriu and Petrea, 2025). Furthermore, AI is leveraged to promote self-directed learning among postgraduate students, support curriculum development, and enhance adaptive and digital learning platforms (Almulla and Ibrahim Ali, 2024; Alnsour et al., 2025).

AI has demonstrated a notable interactive presence among postgraduate students through the use of its tools in developing research ideas, performing spelling and grammar checks, paraphrasing, ensuring textual coherence, and identifying prior studies and relevant literature related to their research topics. This has contributed to the expanded adoption of these technologies within higher education institutions, alongside their role in enhancing academic writing and data analysis processes. Such developments have been reflected in increased research productivity and improved academic performance efficiency (Anani et al., 2025; Vincent et al., 2025).

Despite the multiple positive contributions of AI in supporting the educational process—particularly in personalized learning, improving academic outcomes, and enhancing student engagement—its use simultaneously raises a range of challenges related to output reliability and accuracy, as well as emerging ethical concerns in the context of higher education (Essien et al., 2024; Vieriu and Petrea, 2025; Gammoh, 2024). Excessive reliance on AI tools has been associated with increased academic misconduct, deterioration of intellectual originality, and a decline in the quality of university assignments, which may ultimately threaten graduates’ workforce readiness and undermine employers’ confidence in university outputs (Al Kaabi, 2025). Al Kaabi (2025) examined the impact of university students’ reliance on artificial intelligence tools in completing academic tasks on critical thinking skills, academic integrity, and employability readiness in higher education institutions in the United Arab Emirates. Using a social survey methodology and a stratified random sample of undergraduate students from higher technical colleges, the findings indicated that the majority of students relied heavily on AI tools for academic purposes. This reliance was linked to an erosion of critical thinking and language skills, while senior students demonstrated more advanced and frequent use of AI tools compared to students in earlier academic years. In this context, a global survey across 16 countries revealed that 86% of students regularly use AI in their studies, with 54% using it weekly and over 66% employing it for information retrieval. Furthermore, 80% reported that AI in universities does not fully meet their expectations, 60% expressed concerns regarding data privacy, security, content reliability, and fairness of AI-based assessments, and 50% indicated that excessive reliance on AI would negatively affect their academic performance and reduce the value of their education (Digital Education Council, 2024).

Recent international literature has increasingly examined the relationship between the use of artificial intelligence (AI) in higher education and students’ critical thinking skills, reporting mixed and sometimes contradictory findings (Premkumar et al., 2024; Fischer et al., 2024; Kasneci et al., 2023; Cotton et al., 2023). While some studies suggest that AI-supported learning environments can enhance analytical reasoning and higher-order cognitive engagement by reducing cognitive load and supporting problem-solving processes, others warn that excessive or unregulated reliance on generative AI tools may weaken students’ evaluative judgment, originality, and depth of reasoning (Premkumar et al., 2024; Fischer et al., 2024).

Within this debate, academic integrity emerges as a critical contextual factor that can shape how students interact with AI technologies, influencing whether AI use supports or undermines meaningful learning and critical thinking development (García-López and Trujillo-Liñán, 2025; Perkins, 2023). Academic integrity encompasses honesty, trust, fairness, responsibility, and ethical conduct, and determines whether students approach AI as a supportive learning tool or as a means to circumvent effort and academic rules (Mattar, 2022; Khatri and Karki, 2023). Students with high academic integrity are more likely to use AI responsibly, leveraging it to enhance analytical reasoning, understanding, and creativity. Conversely, students with low integrity may misuse AI, leading to plagiarism, reduced originality, and weakened critical thinking skills (Al Kaabi, 2025; Vieriu and Petrea, 2025; Uddin and Abu, 2024).

Despite its potential influence, limited empirical research has examined academic integrity as a moderating variable that conditions the impact of AI use on critical thinking, particularly among postgraduate students in non-Western higher education contexts (García-López and Trujillo-Liñán, 2025; Perkins, 2023). This gap justifies treating academic integrity as a moderator: it provides the conceptual rationale that the effectiveness of AI in fostering critical thinking is not uniform but depends on students’ ethical engagement with AI tools. In other words, the relationship between AI usage and critical thinking may be stronger or more positive for students with high academic integrity, and weaker or even negative for students with low integrity.

Understanding how students engage with AI tools, and how academic integrity shapes this engagement, benefits from a theoretical perspective that considers attitudes, social expectations, and perceived behavioral control as key influences on academic behavior. In this regard, the Theory of Planned Behavior (TPB), developed by Ajzen (1991, 2002), offers a well-established conceptual lens for interpreting students’ orientations toward the use of artificial intelligence (AI) in academic writing contexts.

According to TPB, individuals’ engagement with a given behavior is informed by beliefs about its expected outcomes, which shape attitudes toward that behavior. Applied to AI-supported academic work, students’ perceptions of the usefulness of AI tools—such as assisting with idea generation, enhancing writing coherence, or facilitating research processes—may influence their propensity to incorporate these tools into their academic practices. At the same time, normative beliefs derived from peers, instructors, and institutional expectations contribute to subjective norms that may either encourage or constrain AI use, particularly when ethical considerations and academic integrity are salient. Perceived behavioral control further reflects students’ sense of capability in using AI effectively and responsibly, taking into account both individual competence and contextual constraints. Higher levels of perceived control are likely to support more deliberate and reflective engagement with AI tools, including attention to originality and critical evaluation of AI-generated outputs. Conversely, limited perceived control may be associated with more superficial or uncritical reliance on AI, potentially affecting higher-order cognitive processes such as analysis, evaluation, and creativity. From this perspective, TPB highlights how the interaction of attitudes, subjective norms, and perceived behavioral control can inform patterns of AI engagement in academic settings, underscoring the importance of pedagogical approaches that promote ethical awareness and critical thinking (Ajzen, 1991, 2002).

Recent research in higher education contexts have indicated that attitudes toward AI, perceived social expectations, and perceived behavioral control remain influential factors in shaping intentions to use emerging educational technologies, including generative AI tools. Empirical studies drawing on TPB-related constructs have shown that positive attitudes toward AI, supportive academic norms, and higher perceived control are associated with greater willingness to adopt AI for academic purposes, as well as more responsible and ethical patterns of use (Ivanov et al., 2024; Ng and Ridzuan, 2025; Ngo et al., 2025). Collectively, these findings support the relevance of TPB as a theoretical reference point for interpreting relationships among AI usage, academic integrity, and critical thinking within contemporary educational environments.

Within the context of higher education in Jordan, institutions face increasing challenges due to the widespread use of artificial intelligence (AI) tools by students. Key issues include rising rates of academic plagiarism, a decline in the originality of submitted work, excessive reliance on these technologies, weakened critical thinking skills, and diminished quality of academic outputs. This situation has placed growing pressure on Jordanian universities to address these risks by developing strategies to curb unethical and irresponsible AI use, adopting advanced plagiarism detection software, enforcing disciplinary measures against students involved in academic dishonesty, and implementing awareness programs highlighting the impact of AI on students’ cognitive skills (Gammoh, 2024).

At the faculty level, nearly half of academic staff (approximately 47%) report using AI tools for certain research and teaching tasks, while fewer than one-third possess sufficient knowledge of the ethical guidelines governing such use; around 70% expressed ethical concerns regarding plagiarism and exam misconduct (Alnsour et al., 2025). Among students, AI adoption is rapidly expanding, with roughly 90% reporting significant use of ChatGPT and a similar proportion employing AI tools for specific academic purposes. Although most students (about 87%) perceive these tools as educationally beneficial, the actual positive impact is more variable: approximately 58% reported improved academic performance, while around 40% indicated significant gains in comprehension (Al Mashagbeh et al., 2025). These findings underscore the necessity of implementing strict data privacy safeguards and adopting a comprehensive, well-considered approach to ensure ethical deployment of AI technologies within educational institutions (Ayasrah, 2025).

Critical thinking involves core cognitive processes, including analysis, evaluation, and reasoning, and is central to achieving logical conclusions and problem-solving. It relies on metacognitive skills such as attention to accuracy, relevance, and depth, and is essential for preparing students to be ethically responsible contributors to society (Dwyer et al., 2014; Bangun and Pragholapati, 2021; Alwehaibi, 2012). Generative AI (GenAI) can support critical thinking by assisting with lower-order tasks like proofreading and initial idea generation, thereby reducing cognitive load and allowing focus on higher-order analytical and creative work. However, excessive reliance may encourage superficial engagement and cognitive offloading, weakening deep cognitive involvement and critical evaluation (Premkumar et al., 2024; Fischer et al., 2024; Shukor and Osman, 2025; Essien et al., 2024).

Among postgraduate students, this dual effect is pronounced: AI may facilitate task completion but can also reduce evaluative reasoning and creativity, potentially fostering cognitive laziness and undermining adherence to ethical academic writing (Nguyen, 2025; Gonsalves, 2024). Irresponsible use threatens the quality of learning, diminishing language, analytical, and evaluative skills, and negatively impacting graduation readiness, research competence, and academic integrity (Al Kaabi, 2025; Vieriu and Petrea, 2025). Academic integrity is an ethical commitment encompassing honesty, trust, respect, fairness, responsibility, and courage, expected from both students and faculty (Mattar, 2022).

The proliferation of AI tools raises ethical concerns, including plagiarism, superficial understanding, bias, copyright issues, and diminished critical engagement, underscoring the need for responsible educational practices (Uddin and Abu, 2024). The threat to academic integrity is especially pronounced when AI-generated texts closely resemble human-authored work, facilitating plagiarism and misconduct, particularly if students fail to disclose AI usage (Khatri and Karki, 2023).

Regulated AI use, combined with measures to curb overreliance, can enhance critical thinking and creativity while respecting research ethics and data privacy. Universities have responded by implementing plagiarism detection tools such as Grammarly and Turnitin, establishing clear policies, and providing specialized training, aiming to balance AI utilization with academic integrity (Al Kaabi, 2025; Alnsour et al., 2025; Khatri and Karki, 2023). Despite the rapid adoption of generative AI in higher education, particularly for research and academic writing, knowledge about its long-term effects on higher-order thinking skills remains limited. Most studies focus on undergraduate students or general academic performance, without addressing advanced skills such as analysis, evaluation, and critical reasoning, especially among postgraduate students expected to demonstrate high levels of critical thinking and adherence to academic integrity (Kasneci et al., 2023; Cotton et al., 2023).

Accordingly, the primary gap addressed by this study is the absence of explanatory models integrating generative AI use with academic integrity as a moderating variable that shapes the relationship between AI utilization and the development of critical thinking skills. Current literature tends to treat academic integrity either as an independent ethical issue or as a potential risk associated with AI misuse, without examining its interactive role in influencing deep cognitive processes among postgraduate students (García-López and Trujillo-Liñán, 2025; Perkins, 2023). This study aims to investigate how the use of generative AI tools influences critical thinking skills among postgraduate students, and to examine the moderating role of academic integrity in this relationship. The study focuses on a context that has been underrepresented in the international literature, specifically postgraduate students in an Arab higher education setting.

The study findings reveal nuanced relationships between AI usage and critical thinking skills among postgraduate students. Lower-order cognitive skills—remembering, understanding, and applying—benefit consistently from AI usage, whereas higher-order skills such as analyzing, evaluating, and creating show weaker direct associations. This pattern suggests that while AI tools facilitate initial stages of knowledge acquisition, their impact on more complex cognitive processes may be limited unless accompanied by ethical and deliberate engagement. Importantly, AI usage dimensions were negatively associated with academic integrity dimensions, particularly with awareness of misuse risks and support for academic writing, highlighting the potential for responsible AI engagement to amplify cognitive gains. These observations underscore the moderating role of academic integrity: students adhering to ethical standards can leverage AI tools to achieve broader cognitive benefits, whereas those with lower integrity may experience limited or uneven advantages. The interpretation of these patterns aligns with Bloom’s Taxonomy, situating AI’s contribution within a hierarchical framework of cognitive development. AI facilitates lower-order cognitive processes, enabling students to allocate more cognitive resources toward higher-order analytical and creative tasks, provided that ethical considerations and responsible use guide their engagement. In this sense, academic integrity functions as a crucial contextual factor shaping the depth and quality of critical thinking outcomes. Thus, rather than repeating the raw correlation values presented in the Results, the discussion emphasizes interpretation, conceptual framing, and the role of academic integrity as a moderator, setting the stage for subsequent sections that analyze detailed paths, latent interactions, and implications for practice and policy.

Among postgraduate students, the use of AI tools is recurrent in tasks such as idea development, spell-checking, paraphrasing, improving coherence, as well as tools that assist in organizing related academic papers. Despite their significance, student’s express concerns about the potential negative effects on critical thinking and creativity, and their intention to use these tools is strongly associated with personal standards and influenced by the surrounding environment (Anani et al., 2025). The AI revolution and its technologies have transformed higher education, placing a significant burden on institutions to address challenges and adopt all possible strategies to enhance their performance in order to remain competitive Amid this revolution, higher education institutions face challenges in effectively fostering students’ critical thinking skills. Despite widespread recognition of its importance for success in the labor market, a gap persists between the acknowledged value of critical thinking and the actual proficiency levels of university students (Hamail, 2023; Premkumar et al., 2024; Zoghbar, 2025).

Several international and regional studies provide additional context for these findings. Al Kaabi (2025) revealed that more than 75% of students used AI programs for academic purposes, and 94.5% of them exhibited deterioration in critical thinking and language skills due to this reliance. Similarly, Aldossary et al. (2024) reported that students hold positive acceptance and high perceptions regarding the expected value of generative AI in education. Despite recognizing the limitations of these tools, students showed a lack of awareness regarding the importance of ethical compliance when using them. The study also indicated that generative AI has a positive impact on achieving sustainable development goals in education. In Saudi Arabia, Almulla and Ibrahim Ali (2024) found that postgraduate students used ChatGPT for various academic purposes, including assistance with homework and research. About 60% of students reported positive effects on their academic performance and research skills. The study also revealed a significant relationship between the frequency of ChatGPT use, students’ satisfaction and perceived usefulness, and its effectiveness in enhancing learning experiences, while also noting a negative impact on critical thinking skills. In the Jordanian academic context, Ayasrah (2025) identified a strong correlation between ethical concerns regarding AI integration and data privacy precautions. The study emphasized the need to establish comprehensive ethical standards to address issues such as algorithmic transparency, bias reduction, and accountability measures. It also highlighted the importance of AI systems in education being transparent and interpretable, and of evaluating these systems in ways that enhance inclusivity and protect data privacy variables, including encryption and cybersecurity.

To contextualize the observed effects of AI usage on cognitive skills, Bloom’s Taxonomy, developed by Benjamin Bloom and colleagues (Bloom et al., 1956), classifies cognitive skills hierarchically, ranging from lower-order skills such as remembering and understanding to higher-order skills including analyzing, evaluating, and creating. This framework is widely used in educational research to design assessments and evaluate the depth of learning outcomes. In line with this framework, the path analysis in the current study revealed that AI usage directly enhances lower-order cognitive skills—remembering, understanding, and applying—while its direct effects on higher-order skills, such as analyzing, evaluating, and creating, are comparatively weaker.

Importantly, the latent interaction analysis demonstrated that academic integrity serves as a significant moderating factor. Students with higher academic integrity were able to leverage AI usage more effectively across all dimensions achieving stronger gains in higher-order critical thinking skills. In contrast, students with lower integrity primarily benefited in lower-order skills, with minimal improvement in higher-order skills. These findings indicate that responsible and ethical use of AI tools can support students’ progression across Bloom’s hierarchical levels, from basic knowledge acquisition to complex analytical and creative thinking. This aligns with previous research showing that structured and guided AI integration enhances both foundational and advanced cognitive skills (Essien et al., 2024).

In a study conducted by Gonsalves (2024) aimed at examining how the integration of AI tools into marketing education assessments affects the critical thinking skills of master’s students in marketing, the results showed that students interacted with AI across Bloom’s cognitive, affective, and metacognitive domains, as well as analytical skills. Bloom’s Taxonomy was proposed as a framework for AI-enhanced learning, encompassing nonlinear learning, strategic thinking, and ethical reasoning. The Kosovo study provides valuable insights into students’ AI use in higher education. It showed that greater engagement with AI tools is associated with more positive attitudes and responsible usage, while lack of training or familiarity with institutional policies creates challenges. These findings support the idea that academic integrity can moderate the relationship between critical thinking and AI use, as adherence to ethical standards and guidelines enhances responsible adoption. The study also indicated that demographic factors had little influence, highlighting the stronger role of experience, social norms, and institutional support (Çerkini et al., 2025).

Moreover, the present findings provide a nuanced understanding of how generative AI usage relates to critical thinking skills among postgraduate students. Path analysis revealed that AI usage directly enhances lower-order cognitive skills—remembering, understanding, and applying—while its direct effects on higher-order skills such as analyzing, evaluating, and creating are comparatively weaker. Importantly, latent interaction analysis demonstrated that academic integrity serves as a significant moderating factor: the positive effects of AI usage on all critical thinking dimensions were substantially stronger for students with higher integrity compared to those with lower integrity. For instance, Academic Support had a pronounced impact on higher-order skills among high-integrity students, whereas its effect was negligible or slightly negative for low-integrity students. Similar patterns emerged for Academic Support and Academic Skills (Learning and Study Aid), highlighting that ethical engagement with AI amplifies cognitive benefits while mitigating potential negative consequences of excessive reliance. Conceptually, academic integrity can be understood as a moderator that shapes the extent to which AI usage translates into meaningful cognitive gains. Students with high integrity are more likely to engage deeply with AI outputs, critically evaluating and integrating information, whereas those with lower integrity may rely excessively on AI, resulting in superficial engagement and reduced critical thinking performance (Uddin and Abu, 2024; Khatri and Karki, 2023; Kasneci et al., 2023). According to Kalniņa et al. (2024) the Low adoption of AI was observed among pre-service teachers, despite recognition of its potential benefits, including language support and access to global knowledge. Challenges such as reduced critical thinking and concerns over plagiarism were also highlighted. These results align with the notion of academic integrity as a moderating factor: students’ adherence to ethical standards and institutional guidelines may strengthen the positive effects of critical thinking on responsible AI use. Furthermore, demographic variables had little influence on AI adoption, emphasizing that experience, social norms, and institutional support are more decisive.

Thus, the moderating role of academic integrity indicates that the benefits or risks of AI usage depend on students’ adherence to ethical standards, emphasizing the need for interventions—such as ethical training, monitoring, and institutional guidelines—to ensure responsible AI integration. This perspective provides a theoretical and empirical basis for understanding differential effects of AI on critical thinking in postgraduate education, addressing gaps in current research that largely focus on undergraduate populations or general academic performance (Cotton et al., 2023).

Finally, our findings have several important implications for future research, educational practice, and policy in higher education. The results suggest that integrating AI tools into postgraduate learning can enhance both lower-order and higher-order cognitive skills when accompanied by ethical guidance and strong academic integrity. Educational institutions should consider designing structured interventions, including ethical training, monitoring, and clear guidelines for responsible AI use, to maximize the cognitive benefits and minimize potential risks such as superficial engagement or reduced critical thinking. These findings also point to the potential for AI-enhanced learning frameworks to be applied across diverse disciplines, promoting strategic, metacognitive, and creative thinking among students.

However, several limitations should be acknowledged. First, the study employed a cross-sectional design, which prevents causal inferences and only allows for associations to be observed. Second, the sample was selected from a single public university and included both master’s and doctoral students. While this approach ensured sufficient statistical power for structural equation modeling, it may limit the generalizability of the findings to other higher education contexts, and future research could examine these relationships separately for master’s and doctoral students to provide more nuanced insights. Third, reliance on self-reported measures may introduce biases such as social desirability or inaccurate reporting of AI usage and critical thinking skills. Future research could also address these limitations by employing longitudinal or experimental designs, expanding samples to multiple universities, and incorporating objective measures of AI engagement and cognitive performance. Despite these limitations, the present study provides a comprehensive understanding of how AI usage and academic integrity interact to influence critical thinking, offering a foundation for both theoretical development and practical applications in higher education.

This study examined how AI usage influences critical thinking skills among postgraduate students, highlighting the moderating role of academic integrity. Results show that AI tools effectively enhance lower-order cognitive skills (remembering, understanding, applying), while higher-order skills (analyzing, evaluating, creating) benefit more selectively. Academic integrity plays a crucial moderating role: students with higher integrity leverage AI to achieve stronger gains across all critical thinking dimensions, whereas students with lower integrity show limited improvement, mainly in lower-order skills. These findings underscore the importance of ethical guidance and institutional measures—such as structured training, monitoring, and clear usage guidelines—to maximize the benefits of AI integration while mitigating potential risks of overreliance. Despite limitations related to the cross-sectional design, self-reported measures, and a single-university sample, the study provides practical and theoretical insights into the interaction between AI usage and academic integrity in shaping cognitive outcomes. Future research should explore longitudinal designs, diverse populations, and objective measures of AI engagement and learning performance.