Artificial Intelligence in Higher Education: A State-of-the-Art Overview of Pedagogical Integrity, Artificial Intelligence Literacy, and Policy Integration

Probably not a substantial claim: may not pass S3 (load-bearing).

Hazari and Vashishth et al. highlight the need for dedicated curricula focused on AI literacy to equip students with a robust understanding of both the capabilities and limitations of AI technologies.

“Hazari [11] and Vashishth et al. [12] highlight the need for dedicated curricula focused on AI literacy, aiming to equip students with a robust understanding of both the capabilities and limitations of such technologies.”

• cites [11] Hazari, 2024, Justification and Roadmap for Artificial Intelligence (AI) Literacy Courses in Higher Education, J. Educ. Res. Pract.Not checked yet

  Not checked in the free run.

• cites [12] Vashishth, T.K., Sharma, V., Sharma, K.K., and Kumar, B. (2024). Enhancing Literacy Education in Higher Institutions with AI: Opportunities and Challenges. Advances in Educational Technologies and Instructional Design, IGI Global. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI can automate content generation, grading assistance, and administrative workflow optimization to reduce faculty workload.

“automated content generation, grading assistance, and administrative workflow optimization that reduce faculty workload”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Jarrah et al. examine the complex relationship between ChatGPT and plagiarism, urging the implementation of strict, transparent policies on the use of generative AI in academic work.

“Jarrah et al. [14] examine the complex relationship between ChatGPT and plagiarism, urging the implementation of strict, transparent policies on the use of generative AI in academic work.”

• cites [14] Jarrah, 2023, Using ChatGPT in Academic Writing Is (Not) a Form of Plagiarism: What Does the Literature Say?, Online J. Commun. Media Technol. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

AI, particularly Generative AI and Large Language Models, is rapidly reshaping higher education by transforming teaching, learning, assessment, research, and institutional management.

“Artificial Intelligence (AI), particularly Generative AI (GenAI) and Large Language Models (LLMs), is rapidly reshaping higher education by transforming teaching, learning, assessment, research, and institutional management.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Ganjavi et al. emphasize the urgent need for academic journals to develop explicit policies on the ethical use of GenAI in scholarly writing, advocating for greater transparency in AI-assisted authorship.

“Ganjavi et al. [10] emphasize the urgent need for academic journals to develop explicit policies on the ethical use of GenAI in scholarly writing, advocating for greater transparency in AI-assisted authorship.”

• cites [10] Ganjavi, 2024, Publishers’ and Journals’ Instructions to Authors on Use of Generative AI in Academic and Scientific Publishing: Bibliometric Analysis, BMJ sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

As early as the 1970s, AI-based systems were first implemented in support of personalized and adaptive instruction, giving rise to the field of Intelligent Tutoring Systems.

“As early as the 1970s, AI-based systems were first implemented in support of personalized and adaptive instruction, giving rise to the field of ITS [17][18][19].”

• cites [17] Conati, 2002, Using Bayesian Networks to Manage Uncertainty in Student Modeling, User Model. User-Adap. Interact. sourceNot checked yet

  Not checked in the free run.

• cites [18] Conati, 2009, Empirically Building and Evaluating a Probabilistic Model of User Affect, User Model. User-Adap. Interact. sourceNot checked yet

  Not checked in the free run.

• cites [19] Luckin, R., and Holmes, W. (2016). Intelligence Unleashed: An Argument for AI in Education, UCL Knowledge Lab.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Studies report that AI improves learning performance and critical thinking.

“Although several studies, e.g., [7,8], report that AI improves learning performance and critical thinking”

• cites [7] Dempere, J., Modugu, K., Hesham, A., and Ramasamy, L.K. (2023). The Impact of ChatGPT on Higher Education. Front. Educ., 8. sourceNot checked yet

  Not checked in the free run.

• cites [8] Wang, 2025, The Effect of ChatGPT on Students’ Learning Performance, Learning Perception, and Higher-Order Thinking: Insights from a Meta-Analysis. Humanit, Soc. Sci. Commun.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The use of AI in education spans several decades.

“Although current public discourse is heavily focused on the latest developments in GenAI tools, the use of AI in education spans several decades (Figure 1).”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

AI, when embraced with ethical discernment and educational accountability, holds the potential to foster a more inclusive, efficient, and democratic future for higher education.

“When embraced with ethical discernment and educational accountability, AI holds the potential to foster a more inclusive, efficient, and democratic future for higher education”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

AI technologies support personalized learning pathways, enhance instructional efficiency, and improve academic productivity by facilitating automated grading, adaptive feedback, and academic writing assistance.

“AI technologies now support personalized learning pathways, enhance instructional efficiency, and improve academic productivity by facilitating tasks such as automated grading, adaptive feedback, and academic writing assistance.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The recent introduction of generative AI and LLMs represents a pivotal shift, bringing AI capable of generating coherent academic text, providing real-time feedback, and supporting complex research tasks.

“The recent introduction of generative AI and LLMs represents a pivotal shift, bringing AI capable of generating coherent academic text, providing real-time feedback, and supporting complex research tasks.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Crompton and Burke's review highlights benefits such as personalized learning and enhanced instruction, while also addressing ethical concerns related to academic integrity and plagiarism detection.

“Previously published work, such as Crompton and Burke's [6] review of AI's potential in universities, highlight benefits such as personalized learning and enhanced instruction, while also addressing ethical concerns related to academic integrity and plagiarism detection.”

• cites [6] Crompton, 2023, Artificial Intelligence in Higher Education: The State of the Field, Int. J. Educ. Technol. High. Educ. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Lack of standardized AI policies and fragmented institutional governance highlight the urgent necessity for transparent frameworks that balance technological adoption with academic values.

“The lack of standardized AI policies and fragmented institutional governance highlight the urgent necessity for transparent frameworks that balance technological adoption with academic values.”

• AssumptionRests on an assumption: There is a lack of standardized AI policies and fragmented institutional governance in higher education.

Probably not a substantial claim: may not pass S3 (load-bearing).

Farahani and Ghasmi argue that challenges necessitate well-structured governance strategies that account for the pedagogical, ethical, and social consequences of AI deployment in higher education.

“These challenges, as Farahani and Ghasmi [15] also argue, necessitate well-structured governance strategies that account for the pedagogical, ethical, and social consequences of AI deployment in higher education.”

• cites [15] Farahani, 2024, Artificial Intelligence in Education: A Comprehensive Study, Forum Educ. Stud. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

Widespread adoption of AI tools among students and faculty has created a critical need for AI literacy encompassing technical proficiency, critical evaluation, ethical awareness, and metacognitive engagement with AI-generated content.

“The widespread adoption of AI tools among students and faculty members has created a critical need for AI literacy-encompassing not only technical proficiency but also critical evaluation, ethical awareness, and metacognitive engagement with AI-generated content.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The history of AI in education illustrates a gradual expansion in functionality and capacity.

“Over the decades, the history of AI in education illustrates a gradual expansion in functionality and capacity (Figure 1).”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Early systems such as Intelligent Tutoring Systems primarily offered rule-based, adaptive learning focused on specific knowledge domains.

“Early systems such as Intelligent Tutoring Systems (ITS) primarily offered rule-based, adaptive learning focused on specific knowledge domains.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Cheng et al. propose concrete recommendations for responsible AI usage in academic composition, underscoring the importance of human oversight and critical evaluation of AI-generated content.

“Similarly, Cheng et al. [3] propose concrete recommendations for responsible AI usage in academic composition, underscoring the importance of human oversight and critical evaluation of AI-generated content.”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

AI poses challenges related to digital divides that exacerbate inequities.

“digital divides that exacerbate inequities”

Probably not a substantial claim: may not pass S2 (checkable).

The success of AI in higher education depends on purposeful integration that balances innovation with academic values such as integrity, creativity, and inclusivity.

“its success depends on purposeful integration, balancing innovation with academic values such as integrity, creativity, and inclusivity.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The adoption of AI in higher education has accelerated significantly, with a wide range of applications now actively reshaping the educational landscape.

“The adoption of AI in higher education has accelerated significantly, with a wide range of applications now actively reshaping the educational landscape [4,5].”

• cites [4] O’Donnell, F., Porter, M., and Rinella Fitzgerald, D. (2024). The Role of Artificial Intelligence in Higher Education. Ir. J. Technol. Enhanc. Learn., 8. sourceNot checked yet

  Not checked in the free run.

• cites [5] Shishavan, H.B. (2024). AI in Higher Education. 2024: ASCILITE 2024 Conference Proceedings, The University of Melbourne.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

With advances in machine learning, data availability, and natural language processing, more sophisticated AI tools emerged, enabling personalized learning paths, automated grading, and plagiarism detection.

“With advances in machine learning, data availability, and natural language processing, more sophisticated AI tools emerged, enabling personalized learning paths, automated grading, and plagiarism detection.”

Probably not a substantial claim: may not pass S4 (based).

AI poses challenges to academic integrity, including plagiarism and misuse of generative content.

“AI poses challenges related to academic integrity (e.g., plagiarism and misuse of generative content)”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Ajani et al. call on educational institutions to reform their academic programs by incorporating both theoretical and practical training in AI, ensuring institutional readiness.

“Ajani et al. [13] further call on educational institutions to reform their academic programs by incorporating both theoretical and practical training in AI, ensuring institutional readiness.”

• cites [13] Ajani, 2024, The Emergence of Artificial Intelligence in the Higher Education, Int. J. Res. Bus. Soc. Sci.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

AI poses challenges related to algorithmic bias and data privacy.

“algorithmic bias and data privacy”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

AI-driven analytics can inform curriculum design and early intervention to improve student outcomes.

“AI-driven analytics that inform curriculum design and early intervention to improve student outcomes”

Probably not a substantial claim: may not pass S4 (based).

AI can provide adaptive tutoring and real-time feedback that tailor instruction to individual learning trajectories.

“Key opportunities include the deployment of adaptive tutoring and real-time feedback mechanisms that tailor instruction to individual learning trajectories”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

The public release of ChatGPT in late 2022 brought AI to the forefront of university discussions, as it marked the first time that a powerful, generative AI tool became widely accessible and capable of producing academic-level output in response to general prompts.

“The public release of ChatGPT in late 2022 brought AI to the forefront of university discussions, as it marked the first time that a powerful, generative AI tool became widely accessible and capable of producing academic-level output in response to general prompts.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Jill Watson, an AI teaching assistant implemented at Georgia Tech in 2016, was developed by Professor Ashok Goel using IBM's Watson platform and deployed in an online course forum to automatically answer frequently asked student questions.

“A notable pre-LLM example of AI use in higher education is the case of Jill Watson, an AI teaching assistant implemented at Georgia Tech in 2016.Developed by Professor Ashok Goel using IBM's Watson platform, Jill was deployed in an online course forum to automatically answer frequently asked student questions.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Faculty training is equally critical: educators must be equipped to integrate AI responsibly into their teaching practices, guide student use, and assess its impact on learning and academic culture.

“At the same time, faculty training is equally critical: educators must be equipped to integrate AI responsibly into their teaching practices, guide student use, and assess its impact on learning and academic culture [45,46].”

• cites [45] Kudiabor, 2024, How AI-Powered Science Search Engines Can Speed Up Your Research, NatureNot checked yet

  Not checked in the free run.

• cites [46] Mollick, E., and Mollick, L. (2023). Assigning AI: Seven Approaches for Students, with Prompts. arXiv. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Roschelle et al. observe that AI improves accessibility, flexibility, and active student engagement, especially in mixed and remote learning contexts.

“Similarly, Roschelle et al. [29] observe that AI improves accessibility, flexibility, and active student engagement, especially in mixed and remote learning contexts.”

• cites [29] Roschelle, J., Lester, J., and Fusco, J. (2020). AI and the Future of Learning: Expert Panel Report, Digital Promise. Available online: https://circls.org/reports/ai-report. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Vashishth et al. mention that education should not be limited to tool usage but must enhance students' capacity to assess the accuracy, reliability, and potential biases of AI outputs.

“As Vashishth et al. [12] mention, education should not be limited to tool usage but must enhance students' capacity to assess the accuracy, reliability, and potential biases of AI outputs.”

• cites [12] Vashishth, T.K., Sharma, V., Sharma, K.K., and Kumar, B. (2024). Enhancing Literacy Education in Higher Institutions with AI: Opportunities and Challenges. Advances in Educational Technologies and Instructional Design, IGI Global. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

Tripathi reveals that many undergraduate students possess low levels of AI knowledge, highlighting the urgent need to institutionalize AI literacy courses at all levels of higher education.

“The recent study by Tripathi [40] reveals that many undergraduate students possess low levels of AI knowledge, highlighting the urgent need to institutionalize AI literacy courses at all levels of higher education.”

• cites [40] Tripathi, 2024, Awareness of Artificial Intelligence (AI) among Undergraduate Students, NPRC J. Multidiscip. Res. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

When prompted to retrieve passages from their AI-generated texts, 83% were unable to repeat even a single sentence, signifying a significant deterioration in episodic memory.

“Furthermore, when prompted to retrieve passages from their AI-generated texts, 83% were unable to repeat even a single sentence, signifying a significant deterioration in episodic memory.”

• cites [39] Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.-H., Beresnitzky, A.V., Braunstein, I., and Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task. arXiv.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Farrelly and Baker emphasise that GenAI technologies are altering the functions of students and educators, necessitating the integration of systematic teaching on ethics, transparency, and responsible usage.

“Farrelly and Baker [2] emphasise that GenAI technologies are altering the functions of students and educators, necessitating the integration of systematic teaching on ethics, transparency, and responsible usage.”

• cites [2] Farrelly, T., and Baker, N. (2023). Generative Artificial Intelligence: Implications and Considerations for Higher Education Practice. Educ. Sci., 13. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Artificial Intelligence has become a foundational tool in higher education, influencing a wide range of academic functions-from teaching and learning to writing, research, and student assessment.

“Artificial Intelligence has become a foundational tool in higher education, influencing a wide range of academic functions-from teaching and learning to writing, research, and student assessment.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Students did not realize that Jill was not a human assistant until the end of the semester.

“Remarkably, students did not realize that Jill was not a human assistant until the end of the semester.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Platforms such as Turnitin began integrating early forms of AI, such as paraphrase recognition, to support academic integrity.

“Platforms such as Turnitin began integrating early forms of AI, such as paraphrase recognition, to support academic integrity.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Shi and Xuwei elucidate the use of platforms like Gradescope and EvalAI for automated feedback provision.

“Shi and Xuwei [49] elucidate the use of platforms like Gradescope and EvalAI for automated feedback provision.”

• cites [49] Shi, 2023, Integration of AI with Higher Education Innovation: Reforming Future Educational Directions, Int. J. Sci. Res.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Ganjavi et al. assert that it is essential for students to be instructed on the ethical frameworks regulating the application of AI in academic writing, especially concerning plagiarism.

“Ganjavi et al. [10] assert that it is essential for students to be instructed on the ethical frameworks regulating the application of AI in academic writing, especially concerning plagiarism.”

• cites [10] Ganjavi, 2024, Publishers’ and Journals’ Instructions to Authors on Use of Generative AI in Academic and Scientific Publishing: Bibliometric Analysis, BMJ sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

The shift to widely accessible generative AI necessitates the re-evaluation of multiple educational parameters, from assignment design and assessment strategies to the graduate skills required in a GenAI-informed academic and professional environment.

“This shift necessitates the re-evaluation of multiple educational parameters-from assignment design and assessment strategies to the graduate skills required in a GenAI-informed academic and professional environment [4,6,13,14].”

• cites [4] O’Donnell, F., Porter, M., and Rinella Fitzgerald, D. (2024). The Role of Artificial Intelligence in Higher Education. Ir. J. Technol. Enhanc. Learn., 8. sourceNot checked yet

  Not checked in the free run.

• cites [6] Crompton, 2023, Artificial Intelligence in Higher Education: The State of the Field, Int. J. Educ. Technol. High. Educ. sourceNot checked yet

  Not checked in the free run.

• cites [13] Ajani, 2024, The Emergence of Artificial Intelligence in the Higher Education, Int. J. Res. Bus. Soc. Sci.Not checked yet

  Not checked in the free run.

• cites [14] Jarrah, 2023, Using ChatGPT in Academic Writing Is (Not) a Form of Plagiarism: What Does the Literature Say?, Online J. Commun. Media Technol. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

These syntheses flag data-privacy and algorithmic-bias risks, digital inequities/infrastructure gaps, and the possibility of diminished human interaction, underscoring the need for careful design, educator training, and cross-stakeholder collaboration so that implementations remain aligned with educational values.

“At the same time, these syntheses flag data-privacy and algorithmic-bias risks, digital inequities/infrastructure gaps, and the possibility of diminished human interaction, underscoring the need for careful design, educator training, and cross-stakeholder collaboration so that implementations remain aligned with educational values [47].”

• cites [47] Takona, 2024, AI in Education: Shaping the Future of Teaching and Learning, Int. J. Curr. Educ. Stud. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Rodafinos illustrates the creative and productive affordances of tools like ChatGPT, while also warning of their potential to undermine academic integrity when used uncritically.

“Rodafinos [38] illustrates the creative and productive affordances of tools like ChatGPT, while also warning of their potential to undermine academic integrity when used uncritically.”

• cites [38] Rodafinos, 2025, The Integration of Generative AI Tools in Academic Writing: Implications for Student Research, Soc. Educ. Res. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Crompton and Burke assert that colleges adopting AI technology have attained enhanced learning personalization, more effective control of faculty workload, and superior tracking of academic success.

“Crompton and Burke [6] assert that colleges adopting AI technology have attained enhanced learning personalization, more effective control of faculty workload, and superior tracking of academic success.”

• cites [6] Crompton, 2023, Artificial Intelligence in Higher Education: The State of the Field, Int. J. Educ. Technol. High. Educ. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Recent psychometric work proposes a four-pillar self-regulated learning profile tailored to AI contexts-motivational components, cognitive/metacognitive strategies, time and task management, and environmental/technological self-regulation-with solid construct validity and criterion links to GPA, technology interest, and digital literacy.

“Recent psychometric work proposes a four-pillar self-regulated learning profile tailored to AI contexts-motivational components (intrinsic/extrinsic motivation, selfefficacy), cognitive/metacognitive strategies, time and task management, and environmental/technological self-regulation-with solid construct validity and criterion links to GPA, technology interest, and digital literacy [36].”

• cites [36] Yurt, 2025, The Self-Regulation for AI-Based Learning Scale: Psychometric Properties and Validation, Int. J. Curr. Educ. Stud. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Cheng et al. further emphasise transparency as a fundamental ethical principle in the implementation of instructional AI.

“Cheng et al. [3] further emphasise transparency as a fundamental ethical principle in the implementation of instructional AI.”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The explosion of educational data from e-learning platforms, MOOCs, and educational software gave rise to learning analytics.

“the explosion of educational data from e-learning platforms, MOOCs, and educational software, which gave rise to learning analytics”

Probably not a substantial claim: may not pass S3 (load-bearing).

Intelligent Tutoring Systems significantly enhance the performance of both students and educators.

“while also significantly enhancing the performance of both students and educators [24]”

• cites [24] Spector, J., Merrill, M., and David, M. (2014). Handbook of Research on Educational Communications and Technology, Springer. [3rd ed.]. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

AI offers novel opportunities but also introduces profound pedagogical challenges for universities, necessitating the adoption of innovative approaches to instruction and curriculum design.

“As noted by Roschelle et al. [29], AI offers novel opportunities but also introduces profound pedagogical challenges for universities, necessitating the adoption of innovative approaches to instruction and curriculum design (Figure 2).”

• cites [29] Roschelle, J., Lester, J., and Fusco, J. (2020). AI and the Future of Learning: Expert Panel Report, Digital Promise. Available online: https://circls.org/reports/ai-report. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

In higher education, applications began to proliferate, including course recommendation systems, automated grading tools for programming and mathematics, and advanced plagiarism detection mechanisms.

“In the context of higher education, applications began to proliferate, including course recommendation systems, automated grading tools for programming and mathematics, and advanced plagiarism detection mechanisms.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Gander and Harris advocate the use of diagnostic instruments for measuring AI-related skills and literacy, thus enabling targeted student support.

“To assess the effectiveness of such programs, Gander and Harris [43] advocate the use of diagnostic instruments for measuring AI-related skills and literacy, thus enabling targeted student support.”

• cites [43] Gander, 2024, Understanding AI Literacy for Higher Education Students: Implications for Assessment, AI High. Educ. Symp.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Jacques et al. warn that, in the absence of pedagogical supervision, automated evaluation instruments could compromise profound learning.

“Jacques et al. [50] further warn that, in the absence of pedagogical supervision, these instruments could compromise profound learning.”

• cites [50] Jacques, 2024, A Synthesis of AI in Higher Education: Shaping the Future, J. Behav. Appl. Manag.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

According to Crompton and Song, AI holds substantial potential to transform higher education by enhancing personalized learning and improving instructional accessibility.

“According to Crompton and Song [25], AI holds substantial potential to transform higher education by enhancing personalized learning and improving instructional accessibility (Figure 3).”

• cites [25] Crompton, 2021, The Potential of Artificial Intelligence in Higher Education, Rev. Virtual Univ. Católica NorteNot checked yet

  Not checked in the free run.

May combine two claims.

Although AI can improve performance, its unselective application may undermine cognitive independence and genuine learning.

“Therefore, although AI can improve performance, its unselective application may undermine cognitive independence and genuine learning.”

• cites [39] Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.-H., Beresnitzky, A.V., Braunstein, I., and Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task. arXiv.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Bittle and El Gayar argue that developing AI literacy strategies is essential to safeguarding academic integrity and preventing instances of plagiarism and ethical misconduct.

“In a more systematic review, Bittle and El Gayar [42] argue that developing AI literacy strategies is essential to safeguarding academic integrity and preventing instances of plagiarism and ethical misconduct.”

• cites [42] Bittle, K., and El Gayar, O. (2025). Generative AI and Academic Integrity in Higher Education: A Systematic Review and Research Agenda. Information, 16. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

AI literacy draws on digital, information, media, data, and metaliteracies, yet adds a distinct critical and ethical dimension, emphasizing the ways in which AI technologies shape cognitive processes, creativity, and decision-making.

“Within this multilayered framework, AI literacy draws on all these literacies, yet adds a distinct critical and ethical dimension, emphasizing the ways in which AI technologies shape cognitive processes, creativity, and decision-making [35].”

• cites [35] Touretzky, 2019, Envisioning AI for K-12: What Should Every Child Know about AI?, Proc. AAAI Conf. Artif. Intell.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The 21st century, and especially the post-2010 era, marked a turning point in the broader adoption of AI in educational settings.

“The 21st century, and especially the post-2010 era, marked a turning point in the broader adoption of AI in educational settings.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Particularly since 2020, with the global proliferation of generative AI tools, the use of these technologies has expanded rapidly with minimal delay.

“Particularly since 2020, with the global proliferation of generative AI tools (e.g., ChatGPT, Copilot, Claude, Gemini, etc.), the use of these technologies has expanded rapidly with minimal delay [1].”

• cites [1] Liu, 2023, Summary of ChatGPT-Related Research and Perspective Towards the Future of Large Language Models, Meta-Radiology sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The EDUCAUSE report delineates AI literacy through four interconnected dimensions: comprehending the foundational concepts of AI systems, using AI tools actively, critically assessing AI-generated outputs, and participating ethically in public discourse regarding AI applications.

“The EDUCAUSE [30] report delineates AI literacy through four interconnected dimensions: comprehending the foundational concepts of AI systems, using AI tools actively, critically assessing AI-generated outputs, and participating ethically in public discourse regarding AI applications.”

• cites [30] EDUCAUSE (2024). Defining AI Literacy for Higher Education, EDUCAUSE. Available online: https://www.educause.edu/content/2024/ai-literacy-in-teaching-and-learning/defining-ai-literacy-for-higher-education.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Adopting frameworks such as the ETHICAL AI Framework promotes responsible use, enhances transparency, and upholds academic integrity within an increasingly dynamic technological ecosystem.

“Adopting frameworks such as the ETHICAL AI Framework promotes responsible use, enhances transparency, and upholds academic integrity within an increasingly dynamic technological ecosystem [44].”

• cites [44] Uzun, 2023, ChatGPT and Academic Integrity Concerns: Detecting Artificial Intelligence Generated Content, Lang. Educ. Technol.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Intelligent Tutoring Systems currently are regarded as one of the most promising mechanisms for the future of education.

“Consequently, ITS currently are regarded as one of the most promising mechanisms for the future of education [19].”

• cites [19] Luckin, R., and Holmes, W. (2016). Intelligence Unleashed: An Argument for AI in Education, UCL Knowledge Lab.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Even prior to 2020, several universities had already experimented with AI-based chatbots for student services and with prototype virtual teaching assistants.

“Even prior to 2020, several universities had already experimented with AI-based chatbots for student services (e.g., FAQs about programs and services) and with prototype virtual teaching assistants [1,7].”

• cites [1] Liu, 2023, Summary of ChatGPT-Related Research and Perspective Towards the Future of Large Language Models, Meta-Radiology sourceNot checked yet

  Not checked in the free run.

• cites [7] Dempere, J., Modugu, K., Hesham, A., and Ramasamy, L.K. (2023). The Impact of ChatGPT on Higher Education. Front. Educ., 8. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Recent systematic reviews converge on a balanced picture: AI not only augments teaching methods and individualizes learning (adaptive pathways, timely feedback), but also streamlines administrative work and cultivates creativity and critical thinking when embedded in thoughtful pedagogy.

“Recent systematic reviews converge on a balanced picture: AI not only augments teaching methods and individualizes learning (adaptive pathways, timely feedback), but also streamlines administrative work (e.g., scheduling, communications) and cultivates creativity and critical thinking when embedded in thoughtful pedagogy [47].”

• cites [47] Takona, 2024, AI in Education: Shaping the Future of Teaching and Learning, Int. J. Curr. Educ. Stud. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Research has shown that the instructional effectiveness of Intelligent Tutoring Systems closely approximates that of human instructors.

“Research has shown that the instructional effectiveness of ITS closely approximates that of human instructors [23]”

• cites [23] VanLehn, 2007, When are tutorial dialogues more effective than reading?, Cogn. Sci. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The perceived effectiveness of Jill Watson illustrated the potential of AI to scale up educational support in large-enrollment online courses.

“The perceived effectiveness of this initiative illustrated the potential of AI to scale up educational support in large-enrollment online courses.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Excessive reliance on AI may intensify performance expectations for both students and instructors and may contribute to cognitive fatigue due to continuous assessment pressures.

“Prillaman [63] warns that excessive reliance on AI may intensify performance expectations for both students and instructors and may contribute to cognitive fatigue due to continuous assessment pressures.”

• cites [63] Prillaman, 2024, Is ChatGPT Making Scientists Hyper-Productive? The Highs and Lows of Using AI, Nature sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Overall, AI is redefining how students learn, how instructors teach, and how universities assess knowledge.

“Overall, AI is redefining how students learn, how instructors teach, and how universities assess knowledge.”

May combine two claims.

In STEM fields, AI is typically used to solve structured problems and automate assessment.

“In the STEM fields, for example, AI is typically used to solve structured problems and automate assessment, while in the humanities and social sciences, its use is more reflective and creative-serving as a dialogic partner or a tool for enhancing argumentation and critical analysis [29,58].”

• cites [29] Roschelle, J., Lester, J., and Fusco, J. (2020). AI and the Future of Learning: Expert Panel Report, Digital Promise. Available online: https://circls.org/reports/ai-report. sourceNot checked yet

  Not checked in the free run.

• cites [58] 2025, AI-Based Applications Enhancing Computer Science Teaching in Higher Education, J. Inf. Syst. Eng. Manag.Not checked yet

  Not checked in the free run.

May combine two claims.

Algorithms adapt content delivery to student profiles, learning performance, and individual preferences, enabling differentiated difficulty levels and targeted scaffolding.

“Algorithms adapt content delivery to student profiles, learning performance, and individual preferences, enabling differentiated difficulty levels and targeted scaffolding [53].”

• cites [53] Krishnakumar, M., and Balasubramanian, K. (2024). Effectiveness of AI in Enhancing Quality Higher Education: A Survey Study. Int. J. Multidiscip. Res., 6. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

When students in the LLM condition wrote without AI support in a follow-up session, they showed impaired recall and remained anchored to linguistic patterns previously suggested by the model.

“Moreover, when students in the LLM condition were asked to write without AI support in a follow-up session, they showed impaired recall and remained anchored to linguistic patterns previously suggested by the model.”

• cites [39] Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.-H., Beresnitzky, A.V., Braunstein, I., and Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task. arXiv.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Rodafinos and Cheng et al. also note the risk of depersonalized thinking and formulaic academic expression.

“However, they also note the risk of depersonalized thinking and formulaic academic expression.”

• cites [38] Rodafinos, 2025, The Integration of Generative AI Tools in Academic Writing: Implications for Student Research, Soc. Educ. Res. sourceNot checked yet

  Not checked in the free run.

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The benefits of AI in personalization, autonomy, and performance enhancement are evident.

“Its benefits in terms of personalization, autonomy, and performance enhancement are evident.”

May combine two claims.

The 24/7 availability of personalized learning support enhances student confidence while reducing anxiety often associated with traditional instruction.

“The 24/7 availability of personalized learning support enhances student confidence while reducing anxiety often associated with traditional instruction [59,60].”

• cites [59] Alasgarova, 2024, The Role of Artificial Intelligence in Shaping High School Students’ Motivation, Int. J. Technol. Educ. Sci. sourceNot checked yet

  Not checked in the free run.

• cites [60] Wang, W., and Li, W. (2024). The Impact of AI Usage on University Students’ Willingness for Autonomous Learning. Behav. Sci., 14. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Entities like UNESCO and the European University Association have urged the creation of formal ethical frameworks for AI application in higher education.

“whilst entities like UNESCO and the European University Association (EUA) have urged the creation of formal ethical frameworks for AI application in higher education [2].”

• cites [2] Farrelly, T., and Baker, N. (2023). Generative Artificial Intelligence: Implications and Considerations for Higher Education Practice. Educ. Sci., 13. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Overdependence on AI may weaken students' conceptual depth and mastery of foundational learning skills.

“However, Huang et al. [61] caution that overdependence on AI may weaken students' conceptual depth and mastery of foundational learning skills.”

• cites [61] Huang, X., Zhang, Y., and Lyu, X. (2023, January 7–9). Assessment of the Impact of Artificial Intelligence on College Student Learning Based on the CRITIC Method. Proceedings of the 2023 IEEE International Conference on Education, Applications and Standards of Converging Technologies (EASCT), Lonavla, India. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

As Roschelle et al. note, such tools support experiential learning and enhance student engagement, particularly in blended or distance learning environments.

“As Roschelle et al. [29] note, such tools support experiential learning and enhance student engagement, particularly in blended or distance learning environments.”

• cites [29] Roschelle, J., Lester, J., and Fusco, J. (2020). AI and the Future of Learning: Expert Panel Report, Digital Promise. Available online: https://circls.org/reports/ai-report. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Wang and Fan's meta-analysis found that the use of AI tools improved learning outcomes, enhanced content comprehension, and fostered higher-order cognitive skills such as critical thinking.

“In particular, Wang and Fan's [8] meta-analysis found that the use of AI tools improved learning outcomes, enhanced content comprehension, and fostered higher-order cognitive skills such as critical thinking.”

• cites [8] Wang, 2025, The Effect of ChatGPT on Students’ Learning Performance, Learning Perception, and Higher-Order Thinking: Insights from a Meta-Analysis. Humanit, Soc. Sci. Commun.Not checked yet

  Not checked in the free run.

May combine two claims.

The integration of AI into university teaching is neither a neutral nor a purely technical process.

“The integration of AI into university teaching is neither a neutral nor a purely technical process.”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

ChatGPT interventions diminished cognitive burden.

“Furthermore, these therapies improved emotional and motivational states-such as self-confidence, concentration, and intrinsic motivation-while diminishing cognitive burden.”

• cites [9] Deng, 2025, Does ChatGPT Enhance Student Learning? A Systematic Review and Meta-Analysis of Experimental Studies, Comput. Educ. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Hicks et al., in a more critical perspective, argue that unchecked use of tools like ChatGPT results in linguistic homogenization and superficial argumentation.

“Hicks et al. [52], in a more critical perspective, argue that unchecked use of tools like ChatGPT results in linguistic homogenization and superficial argumentation.”

• cites [52] Hicks, 2024, ChatGPT Is Bullshit, Ethics Inf. Technol. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Interventions utilising ChatGPT positively influenced student academic performance, with a markedly high total impact size.

“Their findings were substantial: interventions utilising ChatGPT positively influenced student academic performance, exhibiting a markedly high total impact size.”

• cites [9] Deng, 2025, Does ChatGPT Enhance Student Learning? A Systematic Review and Meta-Analysis of Experimental Studies, Comput. Educ. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Selecting and using AI tools demands conscious pedagogical judgment, active supervision, and critical reflection.

“As Cheng et al. [3] emphasize, selecting and using AI tools demands conscious pedagogical judgment, active supervision, and critical reflection.”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

LLMs can support learning when embedded within structured pedagogical frameworks.

“Yet, other studies suggest that LLMs can support learning when embedded within structured pedagogical frameworks.”

• cites [55] Kasneci, 2023, ChatGPT for Good? On Opportunities and Challenges of Large Language Models for Education, Learn. Individ. Differ. sourceNot checked yet

  Not checked in the free run.

• cites [56] Sajja, R., Sermet, Y., Cikmaz, M., Cwiertny, D., and Demir, I. (2024). Artificial Intelligence-Enabled Intelligent Assistant for Personalized and Adaptive Learning in Higher Education. Information, 15. sourceNot checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

The absence of causal evidence on ChatGPT's impact on learning highlights the need for controlled experimental studies.

“The absence of causal evidence highlights the need for controlled experimental studies.”

• AssumptionRests on an assumption: Current cross-sectional findings on ChatGPT and academic performance are inconclusive, with some studies associating ChatGPT use with improved academic performance and others indicating potential negative effects.

Probably not a substantial claim: may not pass S3 (load-bearing).

Cheng et al. advocate for the incorporation of AI disclosure statements in scholarly work.

“Cheng et al. [3] advocate for the incorporation of AI disclosure statements in scholarly work”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Actualising the promise of AI in higher education necessitates institutional investment.

“Nonetheless, actualising this promise necessitates institutional investment.”

May combine two claims.

Professors now use AI to generate quizzes, grade written assignments, provide feedback, and tailor content for students with learning differences.

“Professors now use AI to generate quizzes, grade written assignments, provide feedback, and tailor content for students with learning differences [51].”

• cites [51] Crain, 2025, The Advantages and Disadvantages of AI in Higher Education, Bus. Manag. Rev. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Rodafinos and Cheng et al. examine the positive contributions of AI to assignment writing, particularly in structuring, language precision, and statistical support.

“Rodafinos [38] and Cheng et al. [3] examine the positive contributions of AI to assignment writing, particularly in structuring, language precision, and statistical support.”

• cites [38] Rodafinos, 2025, The Integration of Generative AI Tools in Academic Writing: Implications for Student Research, Soc. Educ. Res. sourceNot checked yet

  Not checked in the free run.

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

ChatGPT interventions improved emotional and motivational states such as self-confidence, concentration, and intrinsic motivation.

“Furthermore, these therapies improved emotional and motivational states-such as self-confidence, concentration, and intrinsic motivation-while diminishing cognitive burden.”

• cites [9] Deng, 2025, Does ChatGPT Enhance Student Learning? A Systematic Review and Meta-Analysis of Experimental Studies, Comput. Educ. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

LLMs can enhance personalization in learning provided they augment rather than replace human judgment.

“Likewise, Sajja et al. [56] emphasize that LLMs can enhance personalization in learning, provided they augment-rather than replace-human judgment.”

• cites [56] Sajja, R., Sermet, Y., Cikmaz, M., Cwiertny, D., and Demir, I. (2024). Artificial Intelligence-Enabled Intelligent Assistant for Personalized and Adaptive Learning in Higher Education. Information, 15. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

AI tools contribute to increased intrinsic motivation, student autonomy, and psychological resilience.

“Artificial Intelligence tools also contribute to increased intrinsic motivation, student autonomy, and psychological resilience.”

• cites [59] Alasgarova, 2024, The Role of Artificial Intelligence in Shaping High School Students’ Motivation, Int. J. Technol. Educ. Sci. sourceNot checked yet

  Not checked in the free run.

• cites [60] Wang, W., and Li, W. (2024). The Impact of AI Usage on University Students’ Willingness for Autonomous Learning. Behav. Sci., 14. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Studies have linked such tools to increased student engagement, self-regulated learning, and academic achievement.

“Studies have linked such tools to increased student engagement, self-regulated learning, and academic achievement [8,54].”

• cites [8] Wang, 2025, The Effect of ChatGPT on Students’ Learning Performance, Learning Perception, and Higher-Order Thinking: Insights from a Meta-Analysis. Humanit, Soc. Sci. Commun.Not checked yet

  Not checked in the free run.

• cites [54] Mohamed, 2024, Empowering the Faculty of Education Students: Applying AI’s Potential for Motivating and Enhancing Learning, Innov. High. Educ. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

ChatGPT performs well on standardized tasks but often fails in complex cognitive challenges when not paired with sound pedagogical design.

“A systematic review by Sumbal et al. [62] highlights that while ChatGPT performs well on standardized tasks, it often fails in complex cognitive challenges when not paired with sound pedagogical design.”

• cites [62] Sumbal, 2024, Can ChatGPT-3.5 Pass a Medical Exam? A Systematic Review of ChatGPT’s Performance in Academic Testing, J. Med. Educ. Curric. Dev. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Platforms with learning analytics provide students real-time feedback on mistakes, time spent, and success/failure patterns, thereby enhancing their capacity to reflect on learning.

“Platforms equipped with learning analytics mechanisms provide students with real-time feedback on their mistakes, the time spent on each task, and patterns of success or failure, thereby enhancing their capacity to reflect on their learning processes [57].”

• cites [57] Msambwa, 2025, The Impact of AI on the Personal and Collaborative Learning Environments in Higher Education, Eur. J. Educ. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Intelligent Tutoring Systems enables real-time monitoring of student progress, identification of cognitive gaps, and the design of tailored learning trajectories.

“ITS enables real-time monitoring of student progress, identification of cognitive gaps, and the design of tailored learning trajectories.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

When intentionally integrated, AI tools can enhance educational efficacy by facilitating personalisation, feedback, accessibility, and empathetic teaching.

“When intentionally integrated, AI tools can enhance educational efficacy by facilitating personalisation, feedback, accessibility, and empathetic teaching.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Unchecked use of LLMs in academic writing may ultimately erode core cognitive capacities despite short-term ease.

“Kosmyna et al. [39] thus offer a critical warning: although LLMs may provide short-term ease, their unchecked use in academic writing may ultimately erode core cognitive capacities.”

• cites [39] Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.-H., Beresnitzky, A.V., Braunstein, I., and Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task. arXiv.Not checked yet

  Not checked in the free run.

May combine two claims.

AI tools are not value-free; they embody design choices, implicit biases, and algorithmic constraints.

“AI tools are not valuefree; they embody design choices, implicit biases, and algorithmic constraints.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Empirical research within institutions, coupled with qualitative and quantitative stakeholder feedback, can guide iterative enhancements in policy and practice.

“Empirical research within institutions, coupled with qualitative and quantitative stakeholder feedback, can guide iterative enhancements in policy and practice.”

May combine two claims.

Institutional governance of AI integration remains fragmented, with inconsistent policies and uneven adoption of AI literacy initiatives contributing to uncertainty.

“Policies for the integration of AI at both institutional and governmental levels have gained considerable importance; however, institutional governance remains fragmented, with inconsistent policies and uneven adoption of AI literacy initiatives contributing to uncertainty.”

May combine two claims.

Independent evidence highlights substantial false-positive risks in AI detection tools, often misclassifying human-written text as AI-generated, with disproportionate impact on non-native English writers.

“Nevertheless, independent evidence highlights substantial false-positive risks, often misclassifying human-written text as AI-generated, with disproportionate impact on non-native English writers, raising fairness and equity concerns [83,84].”

• cites [83] Bjelobaba, 2023, Testing of Detection Tools for AI-Generated Text, Int. J. Educ. Integr. sourceNot checked yet

  Not checked in the free run.

• cites [84] Springer Nature (2023). Why AI Writing Detectors Fall Short: Scientific Critique of Detection Tools, Springer White Papers; Springer Nature.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI tools offer personalized learning experiences, streamline assessment, and support academic productivity.

“AI tools offer personalized learning experiences, streamline assessment, and support academic productivity.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Policy considerations encompass the necessity to reform assessment practices.

“Policy considerations also encompass the necessity to reform assessment practices.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Rodzi et al. contend that conventional evaluation methods may become outdated in the age of generative models.

“Rodzi et al. [90] contend that conventional evaluation methods may become outdated in the age of generative models.”

• cites [90] Rodzi, Z., Rahman, A.A., Razali, I.N.B., Nazri, I.S.B.M., and Abd Gani, A.F. (2023, January 18–19). Unraveling the Drivers of Artificial Intelligence (AI) Adoption in Higher Education. Proceedings of the 2023 International Conference on University Teaching and Learning (InCULT) 2023, Shah Alam, Malaysia. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

These assessment platforms automate grading of large volumes of student work and promote transparency, rubric alignment, and human oversight as a core principle.

“These platforms not only automate the grading of large volumes of student work but also promote transparency, rubric alignment, and the maintenance of human oversight as a core principle [76].”

• cites [76] Instructure (2025, August 05). Instructure and OpenAI Announce Global Partnership to Embed AI Learning Experiences Within Canvas. Available online: https://www.instructure.com/press-release/instructure-and-openai-announce-global-partnership-embed-ai-learning-experiences.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Integrating AI without strategic planning may result in overdependence and diminished expectations for authentic thinking and writing.

“Farrelly and Baker [2] caution that integrating AI without strategic planning may result in overdependence and diminished expectations for authentic thinking and writing.”

• cites [2] Farrelly, T., and Baker, N. (2023). Generative Artificial Intelligence: Implications and Considerations for Higher Education Practice. Educ. Sci., 13. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Collectively, literature review AI instruments augment instructors' metacognitive abilities, promote clarity and justification of pedagogical decisions, and cultivate academic information literacy in higher education.

“Collectively, these instruments augment instructors' metacognitive abilities, promote the clarity and justification of pedagogical decisions, and cultivate academic information literacy in higher education.”

May combine two claims.

A lack of awareness regarding acceptable practices, combined with limited training on evaluating AI-generated content, has led to a rise in plagiarism cases.

“A lack of awareness regarding acceptable practices, combined with limited training on evaluating AI-generated content, has led to a rise in plagiarism cases.”

• Also stated in: § 4. Institutional and Policy Adaptation · plain

  “A lack of awareness regarding acceptable practices, combined with limited training on evaluating AI-generated content, has led to a rise in plagiarism cases.”

• cites [44] Uzun, 2023, ChatGPT and Academic Integrity Concerns: Detecting Artificial Intelligence Generated Content, Lang. Educ. Technol.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

UNESCO advocates using AI as a helper rather than a replacement.

“UNESCO [90,91] similarly advocates using AI as a helper rather than a replacement, a stance echoed across institutional recommendations.”

• cites [90] Rodzi, Z., Rahman, A.A., Razali, I.N.B., Nazri, I.S.B.M., and Abd Gani, A.F. (2023, January 18–19). Unraveling the Drivers of Artificial Intelligence (AI) Adoption in Higher Education. Proceedings of the 2023 International Conference on University Teaching and Learning (InCULT) 2023, Shah Alam, Malaysia. sourceNot checked yet

  Not checked in the free run.

• cites [91] UNESCO (2023). Artificial Intelligence Needs Assessment Survey—Africa, UNESCO. Available online: https://www.unesco.org/en/articles/unesco-launches-findings-artificial-intelligence-needs-assessment-survey-africa.Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

A coherent institutional pathway helps reduce variability in AI integration by aligning governance, literacy, curriculum, and data practices.

“A coherent institutional pathway helps reduce this variability by aligning governance, literacy, curriculum, and data practices.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Sallu et al. emphasise the need for investment in infrastructure, technical expertise, and equitable access to AI so that technology becomes a driver of inclusion rather than exclusion.

“Sallu et al. [89] emphasise the need for investment in infrastructure, technical expertise, and equitable access to AI, so that technology becomes a driver of inclusion rather than a mechanism of exclusion.”

• cites [89] Sallu, 2024, Exploration of Artificial Intelligence (AI) Application in Higher Education, J. Comput. Netw. Arch. High Perform. Comput.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data protection and digital sovereignty are essential elements of institutional AI integration.

“Data protection and digital sovereignty are essential elements.”

May combine two claims.

Excessive dependence on unregulated AI and LLMs threatens to standardise academic discourse and restrict critical critique in educational contexts, potentially fostering overreliance and cognitive deficits ('cognitive debt').

“Excessive dependence on unregulated AI and LLMs threatens to standardise academic discourse and restrict critical critique in educational contexts, potentially fostering overreliance and cognitive deficits (”

• cites [39] Kosmyna, N., Hauptmann, E., Yuan, Y.T., Situ, J., Liao, X.-H., Beresnitzky, A.V., Braunstein, I., and Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task. arXiv.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The ability of LLMs to fabricate citations or produce content lacking authentic grounding raises serious questions about learning quality and the reliability of student submissions.

“The ability of LLMs to fabricate citations or produce content lacking authentic grounding raises serious questions about learning quality and the reliability of student submissions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Such strategies reposition AI as a learning ally rather than a means of circumvention.

“Such strategies reposition AI as a learning ally rather than a means of circumvention [46].”

• cites [46] Mollick, E., and Mollick, L. (2023). Assigning AI: Seven Approaches for Students, with Prompts. arXiv. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Inter-university collaboration is a key enabler for aligning AI practices and shaping regulatory frameworks.

“Finally, inter-university collaboration is a key enabler.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Modern AI tools streamline assignment grading, provide precise and prompt feedback, and alleviate administrative load linked to personalized assessment.

“Modern AI tools streamline assignment grading, provide precise and prompt feedback, and alleviate the administrative load typically linked to personalized assessment.”

Probably not a substantial claim: may not pass S3 (load-bearing).

UNESCO and COPE recommendations advocate for global frameworks grounded in ethical AI use, transparency, and reliability.

“At the level of international policy, both the UNESCO report [86] and the COPE [87] recommendations advocate for global frameworks grounded in ethical AI use, transparency, and reliability.”

• cites [86] UNESCO (2025, August 03). Artificial Intelligence and the Future of Teaching and Learning: Insights and Recommendations. Available online: https://unesdoc.unesco.org/ark:/48223/pf0000386351.Not checked yet

  Not checked in the free run.

• cites [87] Committee on Publication Ethics (COPE) (2023). Guidance for Editors: AI Tools and Authorship, COPE. Available online: https://publicationethics.org.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Comparative case studies demonstrate various levels of preparedness: Finland and Singapore have integrated AI into their curricula and national strategies, while Greece and Germany have deficiencies in institutional coordination.

“Comparative case studies demonstrate various levels of preparedness: nations like Finland and Singapore have integrated AI into their curricula and national strategies, while others, such Greece and Germany, have deficiencies in institutional coordination.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Through inter-university cooperation, institutions can advocate education-specific exceptions and safeguards reflecting teaching and research dynamics.

“Through such cooperation, institutions can advocate education-specific exceptions and safeguards that reflect the unique dynamics of teaching and research.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

AI tools contribute to streamlining daily workflows through automation and cognitive augmentation.

“In this context, AI tools contribute to the streamlining of daily workflows through automation and cognitive augmentation.”

Probably not a substantial claim: may not pass S2 (checkable).

The incorporation of AI into higher education represents a profound alteration of pedagogical methods, learning processes, and student-instructor engagement.

“The incorporation of AI into higher education represents not just technological advancement but a profound alteration of pedagogical methods, learning processes, and student-instructor engagement.”

Classified as probably not a substantial claim.

The rising utilisation of AI by students for academic assistance and assignment composition has led numerous universities to explore strategies for identifying illicit AI usage.

“The rising utilisation of AI by students for academic assistance and assignment composition has led numerous universities to explore strategies for identifying illicit AI usage.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Moya et al. propose international guidelines for AI integration, with particular attention to protecting academic integrity in cross-border education settings.

“Likewise, Moya et al. [88] propose international guidelines for AI integration, with particular attention to protecting academic integrity in cross-border education settings.”

• cites [88] Moya, B., Eaton, S.E., Pethrick, H., Hayden, K.A., Brennan, R., Wiens, J., McDermott, B., and Lesage, J. (2024). Academic Integrity and Artificial Intelligence in Higher Education Contexts: A Rapid Scoping Review. Can. Perspect. Acad. Integr., 7. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Artificial Intelligence is not a neutral tool.

“Artificial Intelligence is not a neutral tool.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A key priority for many universities is institutionalization of policies that balance AI adoption with preservation of academic integrity.

“A key priority for many universities is the institutionalization of policies that balance the adoption of AI with the preservation of academic integrity.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI tools facilitate higher student engagement, autonomy, and accessibility, especially in blended and remote learning environments.

“They facilitate higher student engagement, autonomy, and accessibility, especially in blended and remote learning environments.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Modern AI tools offer university instructors the ability to generate structured, differentiated, and interactive teaching content, significantly reducing preparation time.

“Modern AI tools offer university instructors the ability to generate structured, differentiated, and interactive teaching content, significantly reducing preparation time.”

Probably not a substantial claim: may not pass S3 (load-bearing).

These teaching design platforms empower university educators to design and implement high-quality courses with flexibility, efficiency, and a strong student-centered orientation.

“All together, these platforms, as well as other similar tools, empower university educators to design and implement high-quality courses with flexibility, efficiency, and a strong studentcentered orientation.”

May combine two claims.

Most AI policies prohibit listing AI as an author but allow use of AI tools provided their contributions are transparently acknowledged.

“Most policies prohibit listing AI as an author but allow the use of AI tools, provided that their contributions are transparently acknowledged.”

• cites [10] Ganjavi, 2024, Publishers’ and Journals’ Instructions to Authors on Use of Generative AI in Academic and Scientific Publishing: Bibliometric Analysis, BMJ sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

AI may reproduce biases, perpetuate social inequalities, raise fairness issues-especially for non-native speakers-and exacerbate digital divides between resource-rich and vulnerable institutions or student populations.

“It may reproduce biases, perpetuate social inequalities, raise fairness issues-especially for non-native speakers-and exacerbate digital divides between resource-rich and vulnerable institutions or student populations.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

The strategic integration of AI in higher education requires technological readiness, well-developed AI literacy, institutional coherence, ethical awareness, and pedagogical foresight.

“The strategic integration of AI in higher education requires not only technological readiness and well-developed AI literacy (as previously discussed), but also institutional coherence, ethical awareness, and pedagogical foresight.”

May combine two claims.

These organization tools together enhance teachers' metacognitive and administrative autonomy, fostering an organisational model based on information flow, transparency, and accessibility.

“These tools together enhance teachers' metacognitive and administrative autonomy, fostering an organisational model based on information flow, transparency, and accessibility [79].”

• cites [79] Stanford Report (2023). AI-Assisted Productivity Tools in Higher Education: Trends and Recommendations, Stanford University Publications.Not checked yet

  Not checked in the free run.

May combine two claims.

In early childhood and secondary education, using LLMs to generate tailored feedback on student writing significantly improved academic performance and students' engagement.

“In early childhood and secondary education, a study by Meyer et al. [80] demonstrated that using LLMs to generate tailored feedback on student writing significantly improved academic performance and students' engagement.”

• cites [80] Meyer, 2024, Using LLMs to Bring Evidence-Based Feedback into the Classroom: AI-Generated Feedback Increases Secondary Students’ Text Revision, Motivation, and Positive Emotions, Comput. Educ. Artif. Intell. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

When implemented thoughtfully, AI assessment tools do not merely automate grading but actively enhance the learning experience by providing immediate and meaningful feedback.

“When implemented thoughtfully, these tools do not merely automate grading but actively enhance the learning experience by providing immediate and meaningful feedback.”

Classified as probably not a substantial claim. May combine two claims.

The integration of GenAI into higher education has led to substantial shifts in academic practices worldwide, intensifying discourse around usage policies, academic integrity, and ethical guidelines.

“The integration of GenAI into higher education has led to substantial shifts in academic practices worldwide, intensifying the discourse around usage policies, academic integrity, and ethical guidelines.”

May combine two claims.

Pedagogically grounded AI integration into learning and assessment activities can mitigate covert usage while promoting critical thinking.

“Pedagogically grounded AI integration into learning and assessment activities can mitigate covert usage while promoting critical thinking.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Ongoing assessment of AI integration results is crucial.

“Ongoing assessment of AI integration results is crucial.”

Probably not a substantial claim: may not pass S2 (checkable).

Formulating standards for ethical application of AI in higher education is not solely a technical issue but a pedagogical and institutional necessity.

“Formulating standards for the ethical application of AI in higher education is not solely a technical issue; it is a pedagogical and institutional necessity.”

May combine two claims.

The absence of standardized ethical frameworks for AI use in academic contexts complicates effective implementation and risks confusion or misuse.

“Thus, the absence of standardized ethical frameworks for AI use in academic contexts complicates effective implementation and risks confusion or misuse.”

May combine two claims.

The lack of standardization in AI publishing policies poses challenges for authors and underscores the need for harmonized international guidelines.

“This lack of standardization poses challenges for authors and underscores the need for harmonized international guidelines [10].”

• cites [10] Ganjavi, 2024, Publishers’ and Journals’ Instructions to Authors on Use of Generative AI in Academic and Scientific Publishing: Bibliometric Analysis, BMJ sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

In developing countries, implementation of AI policies requires adaptation to local capabilities.

“In developing countries, the implementation of such policies requires adaptation to local capabilities.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

Disclosure and process documentation increase transparency and trust while preserving academic voice and originality.

“Together, disclosure and process documentation increase transparency and trust while preserving academic voice and originality.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

These assessment tools support grading consistency and save time in resource-constrained academic settings.

“These tools support grading consistency and save time in resource-constrained academic settings.”

May combine two claims.

The empirical findings presented in this entry demonstrate that AI, particularly GenAI, influences various aspects of academic life including written expression support, language revision, productivity enhancement, pedagogical models, assessment strategies, and institutional teaching policies.

“The empirical findings presented in this entry demonstrate that AI, particularly in its GenAI, influences various aspects of academic life, including written expression support, language revision, productivity enhancement, pedagogical models, assessment strategies, and institutional teaching policies.”

Probably not a substantial claim: may not pass S2 (checkable).

Together, the recommended measures align innovation with academic integrity, equity, and durable learning value.

“Together, these measures align innovation with academic integrity, equity, and durable learning value.”

• AssumptionRests on an assumption: The cited literature and comparative case studies are representative of the state of AI in higher education.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Artificial intelligence represents a transformative force in higher education, offering powerful tools to enhance learning, teaching, and research.

“Artificial intelligence represents a transformative force in higher education, offering powerful tools to enhance learning, teaching, and research.”

• AssumptionRests on an assumption: The cited literature and comparative case studies are representative of the state of AI in higher education.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The academic community has begun adopting additional preventive measures that promote transparency and accountability.

“The academic community has begun adopting additional preventive measures that promote transparency and accountability.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

If embraced with institutional commitment and educational accountability, AI can facilitate a more inclusive, efficient, and democratic future in education.

“if embraced with institutional commitment and educational accountability, AI can facilitate a more inclusive, efficient, and democratic future in education.”

• AssumptionRests on an assumption: The cited literature and comparative case studies are representative of the state of AI in higher education.
• AssumptionRests on an assumption: The benefits of AI in higher education are apparent.

Classified as probably not a substantial claim. May combine two claims.

The lack of regulatory consensus, the absence of policy frameworks in many countries, and technological asymmetries between students and institutions underscore the need for strategic institutional integration.

“The lack of regulatory consensus, the absence of policy frameworks in many countries, and technological asymmetries between students and institutions underscore the need for strategic institutional integration.”

• AssumptionRests on an assumption: The cited literature and comparative case studies are representative of the state of AI in higher education.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Various evaluation models and tools have been developed to assess the use and impact of AI in higher education, ranging from diagnostic tools for measuring AI literacy to frameworks for monitoring AI's pedagogical effectiveness and ethical implications.

“Various evaluation models and tools have been developed to assess the use and impact of AI in higher education, ranging from diagnostic tools for measuring AI literacy to frameworks for monitoring AI's pedagogical effectiveness and ethical implications.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

The successful, ethical, and strategic incorporation of AI into higher education relies on the choices made by institutions, educators, and learners.

“The successful, ethical, and strategic incorporation into higher education relies on the choices made by institutions, educators, and learners.”

• AssumptionRests on an assumption: The cited literature and comparative case studies are representative of the state of AI in higher education.

Probably not a substantial claim: may not pass S3 (load-bearing).

Uzun links unregulated AI use to an increase in academic misconduct.

“Uzun[44] links unregulated AI use to an increase in academic misconduct.”

• cites [44] Uzun, 2023, ChatGPT and Academic Integrity Concerns: Detecting Artificial Intelligence Generated Content, Lang. Educ. Technol.Not checked yet

  Not checked in the free run.

May combine two claims.

Ganjavi et al. highlight the absence of standardized regulations for AI utilization in academic production, resulting in varying interpretations of permitted activities by instructors and heightening the risk of confusion or misapplication.

“Ganjavi et al. [10] highlight the absence of standardized regulations for AI utilization in academic production, resulting in varying interpretations of permitted activities by instructors and heightening the risk of confusion or misapplication.”

• cites [10] Ganjavi, 2024, Publishers’ and Journals’ Instructions to Authors on Use of Generative AI in Academic and Scientific Publishing: Bibliometric Analysis, BMJ sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

According to Cheng et al., three core principles guide the ethical use of AI in academic writing: human oversight and evaluation; substantial human contribution to content creation; and explicit disclosure of AI involvement.

“According to Cheng et al. [3], three core principles guide the ethical use of AI in academic writing: (1) human oversight and evaluation; (2) substantial human contribution to content creation; and (3) explicit disclosure of AI involvement.”

• cites [3] Cheng, 2025, Artificial Intelligence–Assisted Academic Writing: Recommendations for Ethical Use, Adv. Simul. sourceNot checked yet

  Not checked in the free run.

May combine two claims.

The commercialization of AI technologies and the inequitable distribution of digital resources may exacerbate the marginalization of students from disadvantaged backgrounds, limiting their capacity to engage fully in educational activities.

“The commercialization of AI technologies and the inequitable distribution of digital resources may exacerbate the marginalization of students from disadvantaged backgrounds, so limiting their capacity to engage fully in educational activities.”