The future of learning or the future of dividing? Exploring the impact of generative artificial intelligence on higher education
Open accessThe DOI matches its Crossref record; 73 of the 73 references it cites were found in a registry or an index; 2 of the 49 sources its claims cite were checked for support.
Abstract
This article examines the impact of generative artificial intelligence (GAI) on higher education, emphasizing its effects in the broader educational contexts. As AI continues to reshape the landscape of teaching and learning, it is imperative for higher education institutions to adapt rapidly to equip graduates for the challenges of a progressively automated global workforce. However, a critical question emerges: will GAI lead to a more inclusive future of learning, or will it deepen existing divides and create a future where educational access and success are increasingly unequal? This study employs both theoretical and empirical approaches to explore the transformative potential of GAI. Drawing upon the literature on AI and education, we establish a framework that categorizes the essential knowledge and skills needed by graduates in the GAI era. This framework includes four key capability sets: AI ethics, AI literacy (focusing on human-replacement technologies), human–AI collaboration (emphasizing human augmentation), and human-distinctive capacities (highlighting unique human intelligence). Our empirical analysis involves scrutinizing GAI policy documents and the core curricula mandated for all graduates across leading Asian universities. Contrary to expectations of a uniform AI-driven educational transformation, our findings expose significant disparities in AI readiness and implementation among these institutions. These disparities, shaped by national and institutional specifics, are likely to exacerbate existing inequalities in educational outcomes, leading to divergent futures for individuals and universities alike in the age of GAI. Thus, this article not only maps the current landscape but also forecasts the widening educational gaps that GAI might engender.
Overview
CoChat Summary
TL;DR
This paper asks whether generative AI in higher education will expand learning or deepen divides, combining a theoretical framework of four AI-era capacities with content analysis of top Asian universities’ GAI policies and core curricula. It finds selective adoption and uneven curriculum integration, which the authors argue risks widening existing educational inequalities.
Key findings
- Among the top 25 Asian universities, the paper’s text reports that only 11 had explicit GAI policies, while Table 1 reports Yes: 12 (48%) and No: 13 (52%). By country/region, China had 0 of 7 (0%) with GAI policies, Hong Kong SAR 5 of 5 (100%), Japan 3 of 4 (75%), Singapore 1 of 2 (50%), South Korea 1 of 5 (20%), Malaysia 1 of 1 (100%), and Taiwan 1 of 1 (100%).
- GAI policies covered areas such as access rights, academic honesty, prompt engineering, awareness, risks and benefits, future of work, ethics and student accountability, value in teaching and learning, defined areas of use, traditional learning and human interaction, target users, decision-making, and contextual/situational approaches; few mentioned full-scale curriculum reforms.
- The paper states that among the top 10 Asian universities with core curricula available in English online, only one university had courses in all categories—computer and digital literacy, AI, human intelligence and capacities, and human-AI collaboration—to equip graduates for the GAI era.
- The theoretical framework identifies four key capability sets: AI ethics, AI literacy (human-replacement technologies), human-AI collaboration (human augmentation), and human-distinctive capacities (unique human intelligence).
Methods
- The study combines a critical literature review on AI, work, and higher education to construct a framework of essential capacities for the GAI era.
- It conducts qualitative content analysis of GAI policy documents and core curricula required for all undergraduates at top Asian universities, defined as Asian universities ranked in the Top 100 of either QS or THE in 2024; data were collected September–November 2024 from publicly accessible English-language university websites.
- Coding was done manually by three coders using keywords and themes guided by the framework, with iterative collaborative discussion until consensus to support intersubjective reliability.
Limitations
- The paper states that institutional heterogeneity could not be explored in depth, including national contexts, enrollment sizes, faculty composition, undergraduate-to-graduate ratios, disciplinary emphases, resources/budgets, and public versus private status; future research should use granular comparative case studies.
- The paper states scope and methodological limitations and suggests future work could use advanced qualitative content analysis, include other regions and graduate-level curricula, and add leading non-Asian institutions such as those in the United States.
- Evident from the text: the reported number of universities with GAI policies is inconsistent—the narrative says 11, while Table 1 totals 12 (48%) and the country table sums to 12; the core curriculum table excerpt also appears incomplete/truncated and does not visibly show the one university said to have all categories.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
2 Valid · 10 Not valid · 54 Not checked · 83 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“Furthermore, much of the existing literature on GAI in education has focused disproportionately on these Western contexts, leaving a gap in understanding how GAI is being integrated into other global regions (Jin et al., 2025).”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > 1. Introduction
What the sources say
- Supportedcites Jin et al., 2025 matched to “Generative AI in higher education: A global perspective of institutional adoption policies and guidelines” (2025) source
The introduction (p. 1) states that 'prior studies have predominantly focused on universities in the Global North ... resulting in findings that may not be representative of a broader, global perspective,' and the Abstract (p. 1) frames the gap as 'most of the prior studies focused on the Global North ... a thorough understanding of the global institutional adoption policy remains absent.' Section 3.1 (p. 4) reinforces this: the sampling was designed 'to provide a comprehensive global perspective and avoid a focus solely on the Global North.' This matches the paraphrased claim, with 'Western contexts' standing in for the source's 'Global North' and the gap in understanding integration in other regions corresponding to the source's non-representativeness point. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“The ability of AI systems to learn from data and make decisions can automate a wide range of tasks, from mundane, repetitive tasks to complex, cognitive tasks (Brynjolfsson et al., 2023).”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > 1. Introduction
What the sources say
- Supportedcites Brynjolfsson et al., 2023 matched to “Generative AI at Work” (2023) source
The Abstract's terminology footnote (p. 1 area) defines machine learning as algorithms that 'learn from data, identify patterns and make predictions or decisions without being explicitly programmed,' which matches the claim's 'learn from data and make decisions' mechanism. Section 1.2 'The Economic Impacts of Generative AI' spans the claim's range: it describes automation of routine work (data entry, bookkeeping, assembly line) and states that recent AI/generative advances allow LLMs to perform a variety of non-routine cognitive tasks (software coding, persuasive writing, graphic design), with Copilot, Elicit and Casetext as examples. The same section and the Abstract add that ML can learn to perform tasks requiring tacit knowledge, i.e. beyond codifiable routine work. The only mild framing difference is that the paper attributes routine-task automation chiefly to earlier computerization while crediting generative AI with extending into non-routine tasks, but the breadth claim itself is clearly treated in our copy. (LLM, conf=0.82)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Contrary to expectations of a uniform AI-driven educational transformation, our findings expose significant disparities in AI readiness and implementation among these institutions.”
- Also stated in: § Abstract · boosted
“Our analysis shows sharp disparities in GAI readiness across top universities”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > Abstract
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Abstract · boosted
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“For instance, some universities allowed only faculty members and certain students enrolled in specific programs to access GAI resources, thus ensuring that these powerful tools are used responsibly.”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > 6. Analysis and findings
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, a significant gap was identified in the absence of fullscale curriculum reforms in the GAI policies, which are affirmed by the examination of the core curricula of some of the top Asian universities.”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > 7. Discussion: selective adoption, equity, and divergence
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on 2 assumptions not shown here.
43 more claims
Create a free account to see all 149 claims and to verify the 31 references still open.
- Probably not a substantial claim. 79
- It may state two separate claims as one. 39
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > 6. Analysis and findings
“out of the 25 top-ranking Asian universities, only 11 had policies explicitly related to GAI.”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > Additional information
“| Total: 25 | | | | Yes: 12 ( 48%); |”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > Abstract
“generative artificial intelligence (GAI)”
§ The future of learning or the future of dividing? Exploring the impact of general artificial intelligence on higher education > Abstract
“general artificial intelligence (GAI)”
Citations
Cites
DiMaggio, P. J., & Powell, W. W. (1983) · American Sociological Review · doi:10.2307/2095101 · Cited by 38,435
DiMaggio, P. J., & Powell, W. W. (2004) · Advances in Strategic Management · doi:10.1016/s0742-3322(00)17011-1 · Cited by 25,802
The future of employment: How susceptible are jobs to computerisation?
Frey, C. B., & Osborne, M. A. (2017) · Technological Forecasting and Social Change · doi:10.1016/j.techfore.2016.08.019 · Cited by 8,727 · cited 1×
“One of the most prominent impacts of AI, and particularly GAI, on the future of work is automation (Frey and Osborne, 2017; Wong, 2020).”
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-019-0171-0 · Cited by 7,242 · cited 1×
“Generative artificial intelligence (GAI), including transformative technologies such as ChatGPT, is rapidly changing the contours of various sectors of human life (Zawacki-Richter et al., 2019; Galindo et al., 2021).”
Novak, J. D., Gowin, D. B., & Kahle, J. B. (1984) · doi:10.1017/cbo9781139173469 · Cited by 5,199
Skills, Tasks and Technologies: Implications for Employment and Earnings
Acemoglu, D., & Autor, D. (2011) · Handbook of Labor Economics · doi:10.1016/s0169-7218(11)02410-5 · Cited by 3,799
Harrison, A., Robinson, C., Ashenfelter, O., & Layard, R. (1989) · The Canadian Journal of Economics · doi:10.2307/135687 · Cited by 3,153
Students' voices on generative ai: perceptions, benefits, and challenges in higher education
Chan, C. K. Y., & Hu, W. (2023) · arXiv · doi:10.48550/arxiv.2305.00290 · Cited by 2,504 · cited 1×
“It also calls for a commitment to lifelong learning, acknowledging that education is a continuous process in the face of rapid technological change (Chan and Hu, 2023; Bowen and Watson, 2024).”
Students’ voices on generative AI: perceptions, benefits, and challenges in higher education
Chan, C. K. Y., & Hu, W. (2023) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-023-00411-8 · Cited by 2,301 · cited 1×
“It also calls for a commitment to lifelong learning, acknowledging that education is a continuous process in the face of rapid technological change (Chan and Hu, 2023; Bowen and Watson, 2024).”
Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence
Noy, S., & Zhang, W. (2023) · doi:10.2139/ssrn.4375283 · Cited by 2,157
Research works printed but not linked: 5
- Heaven, WD (2023) The education of ChatGPT. MIT Technology Review.
- 2023, Future of Jobs Report 2023
- Li, 2023, The Worlds I See
- Bradford, 2023, The Global Battle to Regulate Technology
- Nye, 2004, Soft Power
Cited by
Zeivots, S., Casey, A., Winchester, T., Webster, J., Wang, X., Tan, L., Smeenk, W., Schulte, F. P., Scholkmann, A., Paulovich, B., et al. (2025) · Postdigital Science and Education · doi:10.1007/s42438-025-00595-4 · Cited by 34
Xu, J., Shen, R., Li, J., & Hu, D. (2025) · Australasian Journal of Educational Technology · doi:10.14742/ajet.10656 · Cited by 7
Fu, G., & Cao, K. (2026) · Frontiers in Education · doi:10.3389/feduc.2026.1833616 · Cited by 4
Valentini, A. (2026) · Revista Educación Superior y Sociedad (ESS) · doi:10.54674/ess.v37i2.1278 · Cited by 3
Working memory and generative AI tools in higher education: A systematic review
Almudara, S. B., Alshehri, O., Alzahrani, S. Y., Abdellatif, M. S., Idris, A., & Ibrahim, A. R. (2026) · Research Journal in Advanced Humanities · doi:10.58256/ca66cf85 · Cited by 2
Wong, W., Aristidou, A., & Scheuermann, K. F. (2025) · Data & Policy · doi:10.1017/dap.2025.10023 · Cited by 2
Self-citationThe Cybersociology of Learning as Human-AI Partnerships in the Teaching of Political Education
Yakin, A. M. A., Matari, A. S. A., Espinosa, K. P. M., Enrique, W. D., Cardoso, L., & Muthmainnah, M. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-2600-1458-5.ch006 · Cited by 1
Underachievement Should Be Studied in the School Context
Matthews, M. S. (2026) · Springer eBooks · doi:10.1007/978-3-032-21277-1_6 · Cited by 1
Berroa, A. E. M., & Guerra, Y. H. (2026) · Revista Educación Superior y Sociedad (ESS) · doi:10.54674/ess.v37i2.1102 · Cited by 1
Boichenko, M. A., Zavolodko, H. E., Chernobrovkina, S. V., Antonova, I. V., Karabin, O. Y., Kirdan, O. L., Koval, A. M., Zheliezko, A. M., Popova, M. A., Sbruieva, A. A., et al. (2025) · Izdevniecība “Baltija Publishing” eBooks · doi:10.30525/978-9934-26-600-3
Niu, C., Sadeghi, S., Jin, R., Grote-Garcia, S. A., Love, A., Waltman, B., Garfield, T. A., & Hernández, A. E. (2025) · Intelligent Technologies in Education · doi:10.53761/ssw9kz69
Zhang, Q., & Osborne, C. (2026) · Chinese as a Second Language Research · doi:10.1515/caslar-2026-2002
Abdelsattar, A., Labib, W. A., Abdelalim, M. F., & Abowardah, E. S. (2026) · Frontiers in Education · doi:10.3389/feduc.2026.1763439
Tamam, M. B. (2026) · J-MACC Journal of Management and Accounting · doi:10.52166/j-macc.v9i1.12638
Alvarado, A. L., Espinoza, Ó., Thayer, T., & Chávez, M. E. A. (2026) · Revista Educación Superior y Sociedad (ESS) · doi:10.54674/ess.v37i2.1090
Knowledge structure and future directions in emotional AI tourism research
Mandagi, D. W. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-01723-0
Kanoh, H. (2026) · AHFE international · doi:10.54941/ahfe1008102
Wu, F., Xiao, H., & Lin, X. (2026) · SAGE Open · doi:10.1177/21582440261477182
Governing artificial intelligence use in academia: A virtue-based governance perspective
Chandra, Y., & Liu, G. (2026) · Data & Policy · doi:10.1017/dap.2026.10087 · Cited by 2
Файфура, В. В., & Nasminchuk, I. A. (2026) · Izdevniecība “Baltija Publishing” eBooks · doi:10.30525/978-9934-26-708-6-14
Liu, Y. (2026) · Lecture Notes in Education Psychology and Public Media · doi:10.54254/2753-7048/2026.ld36337
Ferreira, A. F. M. (2026) · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)
Designing for Professional Competence: A Competence Architecture for IS Curriculum Design
Zhou, Y., Guo, Z., Kirshner, S., & Guo, S. (n.d.) · AIS Electronic Library (AISeL)
Ma, J., & Zhang, Y. (2026) · International Conference on Intelligent Education and Computer Technology (IECT) · doi:10.1145/3845354.3845442
Metadata
- Authors
- Wong, W., Aristidou, A., & Scheuermann, K.
- Affiliations
- Chinese University of Hong Kong (HK)
- University College London (GB)
- Corresponding author
- Wong, Wilson
- Venue
- Data & Policy
- Publisher
- Cambridge University Press (CUP)
- Published
- 2025
- Volume
- 7
- ISSN
- 2632-3249 (ISSN-L)
- OpenAlex
- W4411615628
- Semantic Scholar
- 511c582ac546ffcc7230f27f10c82bfa5f8f6968
- Cited by
- 23
- References
- 63
- Citations by year
- 2026: 18 · 2025: 5
- FWCI
- 13.21
- Type
- journal article
- Language
- en
- Update notices
- correction (2025-06-25)
- Open access
- gold
- doi.org
- Locations
- Data & Policy (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- UCL Discovery (University College London) (submittedVersion)
- Funders
- Nanyang Technological University
- Seoul National University
- Chinese University of Hong Kong
- Topics
- Social Sciences › Education › Artificial Intelligence in Education
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Social Sciences › Education › Digital Education and Society
- Keywords
- generative AI, higher education, artificial intelligence in education, educational inequality, AI literacy, AI ethics, human-AI collaboration, human augmentation, graduate competencies, AI policy, higher education curriculum, AI readiness, Asian universities, workplace automation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-11-20, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-11-20 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.
