Comparing the Sustainable Role of Higher Education in National Artificial Intelligence Strategies Through the Lens of Policy Documents in China, Japan, and South Korea (2017–2025)

Abstract

The sustainable role of higher education and its governance logic is increasingly prominent in national artificial intelligence (AI) strategies. Based on core AI-related policy documents issued by China, Japan, and South Korea between 2017 and 2025, the corpus-based analysis conducts a systematic comparison, by means of text coding, from strategic positioning, policy architecture, educational philosophy, and a governance model. All three countries have established AI as a critically sustainable component of national development strategy and explicitly defined its functional role in higher education at the strategic level. However, varied distinct differences are found in the policy implementation approaches and governance structures. China adopts a centralized, state-led approach characterized by a high degree of integration among institutional instruments. Japan reinforces liberal education reform in the context of higher education and extends educational responsibilities across society through a lifelong learning framework. Driven by global technological competition, South Korea advances digital transformation of the education system, while building a trustworthy AI governance system. These divergent policy models reveal the sustainable interplay between national AI strategies and higher education. Overall, our findings indicate that the evolution trajectories of AI strategies affect the stability and adaptive capacity of higher education systems in the long term.

1. Introduction

The rapid advancement of artificial intelligence (AI) has recently exerted transformative effects across all industrial sectors, as witnessed in accelerated reforms in the context of higher education. Accordingly, an increasing number of countries, considering AI as a key component of national strategy, are orienting higher education systems through policies to address the challenges to education reform in the AI era. Specifically, higher education is mainly entrusted with the crucial role of supporting stable innovation capacity at the national level and building sustainable competitiveness.

Early scholarly debates concerning the relationship between AI and higher education focused primarily on teaching practice and learning support. As such, AI applications in this line of research involve intelligent tutoring systems, adaptive learning environments, and learning analytics, underscoring their potential to enhance instructional efficiency and personalized learning [1]. With the advent of Generative Artificial Intelligence (GenAI), discussions have expanded in more depth to curriculum design, learning feedback, and the reconstruction of assessment mechanisms. Essentially, related literature generally holds that GenAI is reshaping modes of learning and the organizational logic of teaching [2]. However, the impact of AI on higher education is not limited to technological applications in classrooms. As Yuan [3] notes, conversations about AI-driven educational transformation often unfold alongside high technological expectations, yet both structural adjustments and governance logic transformations in educational systems are inevitably shaped not by technological development but by policy choices and institutional path dependencies. Therefore, understanding higher education reform in the AI era necessitates a shift from mere quest for teaching applications to in-depth analyses of national policies and institutional frameworks. Furthermore, compared to other emerging issues in higher education, AI occupies a more prominent strategic position within national policy systems. Emerging interdisciplinary fields like educational neuroscience, for instance, though similarly involved in the integration of technology and education, are predominantly incorporated into policy as supports for scientific research or as tools for evidence-based decision-making [4]. In contrast, AI is not only introduced into the higher education system as a technological tool but is also regarded as a crucial component of the national innovation system, and is advanced through sustained policy frameworks.

The implementation effectiveness of policy documents varies significantly across different countries. In Singapore, for instance, which has a strong tradition of centralized governance, policy documents often directly guide the allocation of resources in higher education through legislation or administrative directives. In contrast, in Japan, with its deep-seated tradition of institutional autonomy, national-level policy documents primarily provide strategic visions and developmental directions, and their implementation relies on the autonomous interpretation by individual higher educational institutes. Though sharing East Asian cultural context, China and South Korea exhibit structural differences in their policy models with Japan, which indicates that even under similar AI strategic frameworks, the role and sustainability of higher education may follow distinct pathways. Therefore, when comparing the function of higher education within the AI strategies of these three countries, it is essential to consider the institutional nature of policy documents and their actual effectiveness within each national context.

As illustrated above, while previous studies have established a connection between AI applications and governance model, to what extent this link is differentiated across countries in a particular region remains unaddressed. Accordingly, this study sets out to target national AI strategy documents issued by China, Japan, and South Korea as our research corpus and conduct a structured analysis of the document content associated with higher education. By developing a unified coding framework, it systematically compares the policies of the three countries in terms of strategic positioning, policy instrument application, and sustainable role allocation and governance approach for higher education, thereby revealing differences in their policy priorities and governance logics. In this study, the concept of sustainable role is defined as the institutional arrangements that enable education systems to maintain stability, adaptability, and long-term developmental capacity amid technological change. This conceptualization integrates a dual perspective: higher education as a supportive structure of national innovation systems, and AI as a policy-driven technological transformation, providing a unified framework for cross-national comparative analysis in this paper. Overall, we strive to clearly present the varied approaches through which these three Northeast Asian countries configure the role of higher education in the AI context.

2. Literature Review

2.1. Policy-Oriented Foci on Artificial Intelligence in Higher Education

The application of AI in higher education has evolved from functional technological support to integration into systemic learning environments. Early studies primarily involved intelligent tutoring systems, adaptive learning environments, and learning analytics, emphasizing the enhancement of instructional efficiency and personalized support through algorithmic prediction, data analysis, and automated feedback mechanisms [5].

Taken together, more scholarly attention has been so far given to policy-driven relevance of AI in higher education than use-oriented investigations. Foci on AI in higher education have expanded from a single pedagogical aid to a force shaping learning structures and organizational management. However, much is underexplored inasmuch as research attention gradually shifts from technological evaluations toward the interaction between learning processes and institutional policies. It is, to a significant degree, shaped by the policy environment and is intertwined with the long-term stability and adaptive capacity of educational systems amidst technological change. For instance, universities introducing GenAI often influence policy-makers’ decisions at national level by establishing varied documents, such as academic norms and guidelines of using AI [6,7]. Concurrently, studies show that AI application is subject to significant influence from regulatory frameworks and institutional environments, and variations in national policy systems further shape its specific modes of application and developmental trajectories [8,9], exhibiting distinct characteristics of institutional constraints, particularly in areas like data privacy and governance mechanisms [10]. Furthermore, factors such as data privacy protection, intellectual property regimes, and platform governance rules impose important constraints on how and to what extent AI technologies are applied in educational settings, thereby influencing the sustainable developmental capacity of higher education systems [11,12,13,14]. Therefore, understanding the relationship between AI and higher education solely from an application perspective remains limited. It is necessary to further examine this relationship from policy and institutional dimensions to uncover the sustainable role of higher education in the development of AI. Based on this, the next section will review the governance challenges of AI in higher education.

2.2. Governance Challenges of Generative Artificial Intelligence to Higher Education

The widespread adoption of GenAI poses multifaceted challenges to higher education, ranging from technical adaptation to institutional restructuring. The explosive uptake of GenAI tools, such as ChatGPT and Deepseek, has brought governance issues pertinent to those tools to the forefront of policy debates, since their use is inevitably accompanied by severe concerns, including questionable content accuracy, diminished reliability of assessments, and technological misuse potential [15].

With respect to institutional policy and regulation, GenAI has blurred the boundary between originality and automated generation, thereby challenging academic integrity and evaluation systems. Institutional responses at the higher education level have primarily captured the delineation of responsibilities, transparency of use, and mechanisms for human oversight, attempting to strike a balance between encouraging innovation and maintaining scholarly norms [16,17,18]. Such efforts indicate that higher education institutions are staking out their claim to defining the legitimate boundaries of AI through formalized institutional documents.

As regards competency structure, the emergence of GenAI has substantially advanced the reassessment of talent cultivation objectives and knowledge structures. Notedly, national AI strategies have incorporated education into the framework of capability building, highlighting the cultivation of digital literacy and AI-related competencies through higher education to support national innovation systems and industrial development [19]. In this context, the role of higher education is no longer confined to the transmission of knowledge but redefined as a key carrier for developing innovation and technological capabilities. Consequently, the transformation of competency structure has become a central issue in formulating and implementing AI strategies.

As for ethics and governance, GenAI has spurred extensive discussion concerning fairness, transparency, and attribution of responsibility. Wright [20] notes that inconsistencies may exist between the articulation of AI ethical principles and their institutional implementation, given that the value declarations in policy texts do not necessarily translate fully into concrete institutional arrangements. Moreover, from a global governance perspective, the formulation of AI policy is influenced by international organizations and transnational policy discourses [21]. This technological outreach indicates that AI governance involves not merely technical issues but also value choices and institutional coordination.

Based on the aforementioned three dimensions of leveraging AI, prior studies have further explored the role of AI within a sustainable development framework. Tanveer et al. [22] contend that AI influences not only instructional practices but also the accumulation of educational capital and the long-term development capacity of education systems in developing countries. This suggests that the AI-related policies in the educational field extend beyond a mere technological tool, and thus become increasingly intertwined with the national development paradigm and the sustainable transformation of education systems. Meanwhile other scholars have reflected on the preconditions for applying AI in various education scenarios through the lenses of values and norms. They emphasize that, in the context of ongoing digitalization and AI proliferation, the sustainability of education necessitates a reassessment of its value orientations and developmental objectives [23]. Within this long-term framework, AI policy concerns not only choices of technological approaches but also institutional arrangements regarding educational purposes and governance logic.

The emergence of GenAI poses multiple challenges to higher education governance, including the need for normative reconstruction, capability recalibration, and the institutionalization of ethical oversight. Although existing studies have examined institutional responses within higher education institutions, as well as national strategic frameworks and ethical principles, the issues at stake have increasingly moved beyond mere technological adaptation down to the reconfiguration of institutional logics and governance architectures. Given that GenAI features transnational diffusion and global competition, national policy responses are, more often than not, situated within a broader context of international policy interactions. Prior research has shown that when confronting similar policy challenges and technological issues, different countries may exhibit a certain degree of convergence in policy ideas and institutional rhetoric [24], and may also fine-tune their domestic policy directions by drawing on the experiences of other countries or absorbing initiatives advanced by international organizations [25]. Nevertheless, such similarities do not imply simple replication of institutional structures; rather, they are interwoven with existing governance traditions and development strategies of each country in sustainable ways.

2.3. Research Gaps and Positioning

A synthesis of existing studies, though, indicates that the long-term interactive relationship between AI and higher education has been discussed. Overall, these studies primarily concentrate on either technological applications or governance challenges, and a rare body of investigations specifically targets the enduring functions and institutional roles of higher education in AI development from a national strategic perspective. In particular, from a cross-national comparative viewpoint, some studies have compared AI strategy frameworks across multiple countries [26,27], yet how different countries embed higher education into their AI development pathways through policy systems, and how this embedding process reflects its “sustainable role,” still requires in-depth exploration. Therefore, conducting a systematic comparative analysis of the sustainable role of higher education in AI strategies from the level of national policy constitutes the primary starting point of this study.

Another line of existing international comparative studies exhibits a notable contextual limitation. Their sources of policy texts predominantly come from the English native countries, arguably giving rise to comparatively insufficient in-depth excavation and analysis of original policy documents issued by non-English speaking countries, particularly those in Chinese, Japanese, and similar Asian languages. As a result, interpretations of policy discourses, value hierarchies, and concrete implementation mechanisms of countries, such as China and Japan, often depend on translated or summarized secondary materials. Not unexpectedly, they may fail to accurately capture the original context and nuanced meanings of their source policy texts.

To address this gap, this study aims to answer two critical questions:

RQ1: What differences exist among the national higher education AI policies of China, Japan, and South Korea in terms of policy priorities, competency-structure positioning, and governance logic?

RQ2: How do these differences reflect institutional choices of each country regarding the role of higher education, the construction of competency structures, and governance logic in the era of AI?

To explore the questions above, we construct a unified text-coding framework to perform a structured analysis by comparing the role of higher education in national AI policy documents selected from China, Japan, and South Korea. By coding and comparing policy texts within a common set of dimensions, we attempt to reveal differences in policy architecture and examine how institutional contexts shape governance models for AI in higher education.

3. Research Design

3.1. Subjects

This study selects 15 national AI policy documents issued by China, Japan, and South Korea between 2017 and 2025, encompassing national AI development strategies, basic AI plans, educational reform blueprints, and other policy documents closely related to AI education (

Table 1. National AI education policy documents issued by China, Japan, and South Korea.

Country	Number	Time Span	Issuing Agencies	Policy Level and Type
China	5	2017–2025	The State Council; Ministry of Education	National strategy + Targeted action
Japan	5	2017–2025	Cabinet Office; Ministry of Education, Culture, Sports, Science and Technology (MEXT)	National strategy + Guidance plan
South Korea	5	2019–2024	Ministry of Science and Information and Communications Technology (MSIT); Ministry of Education	National strategy + Implementation plan

). The above national policy documents align the following four selection criteria:

(1)

The documents are issued or led by government departments or central agencies at national level;

(2)

The documents treat AI as a significant policy agenda;

(3)

The documents explicitly underscore education and talent cultivation in their content;

(4)

The documents could represent the overall direction of national education policy rather than local measures.

3.2. Research Methods

This study employs qualitative textual analysis to conduct a structural comparison of AI policies used in China, Japan, and South Korea. Textual analysis is capable of systematically revealing policy frameworks and institutional logics through the classification and interpretation of textual materials [28]. Qualitative content analysis emphasizes the construction of explicit coding categories, though arbitrarily assigned, and closely aligns analytical dimensions with the two research questions to generalize meaning segments inductively from texts, thereby enhancing the transparency and reproducibility of comparative research [29]. These existing studies provide methodological foundations for qualitative analysis and policy text analysis, serving to guide the delineation of analytical units and the standardization of the coding process.

In view of the strategic, institutional, and cross-sectoral characteristics of AI policies, quantitative indicators are insufficient to capture their rich policy content and governance orientations. Therefore, dividing national policy texts into segments for analysis and constructing a unified coding framework, as illustrated in

Table 2. Coding framework for policy text analysis.

Dimension	Operational Definition	Core Characteristics
Policy Positioning (PP)	Delineation of AI development goals, strategic significance, and its relationship to national development within policy texts	Whether AI is defined as a core strategic technology, a foundational social capability, or a tool for global competition
Policy Advancement (PA)	Model of policy evolution from planning to implementation	Phased advancement, progressive integration, or concentrated promotion
Role of Higher Education (RE)	Functions, tasks, and positioning of cultivation models assigned to higher education within the AI policy framework	National innovation system unit, social capability cultivation platform, or technological application carrier
Realization Mechanism for Educational Reform (RM)	Specific approaches for promoting education reform	Academic system building, curriculum standards, project-based approaches, or direct technological embedment
Talent Cultivation Orientation (TO)	Types of talent shaped by policies	Interdisciplinary, liberally educated, or personalized competency-oriented talents
Governance and Risk Response (GR)	Logic for addressing ethics, risk, and social impact	Institution-first, value-led, or trust-building approaches
Policy Scope and Coverage (PS)	Social coverage of AI education	Higher education, lifelong learning, or the entire workforce
International Participation (IP)	Globalization approaches for AI education	Institutional participation, value export, or industrial collaboration

, this study adopts an inductive approach to establish eight analytical dimensions, such as strategic policy positioning of AI, the role of higher education, and talent cultivation orientation. In what follows, the specific analyses for each country and the comparison of policies all align with this framework. The established coding dimensions encompass both descriptive aspects of policy characteristics, such as policy positioning and policy advancement, and analytical aspects of policy orientations and governance mechanisms, such as governance and risk response. Together, they constitute a multidimensional analytical framework for examining national AI policy systems.

This study selected a total of 15 national-level AI policy documents from China, Japan, and South Korea. These documents were divided into 180 AI-themed analytical units for coding: 69 units from China, 55 units from Japan, and 56 units from South Korea. This study treats segments for analysis in policy texts as the basic coding object, and thus selects relatively independent sentences or clauses that can semantically express policy goals or intentions. This approach helps preserve the integrity of policy context while reducing, to some extent, the interference caused by syntactic differences in cross-text and cross-national comparative analyses.

In the specific coding stage, a phased procedure is adopted to enhance the systematicity and transparency for analysis. First, policy documents are read holistically to clarify the content related to general positioning of AI strategic objectives and education. Second, each document is automatically examined by inputting key index terms, such as “artificial intelligence/AI”, “GenAI”, “higher education”, and “governance risk”. Next, both authors manually screen out those texts irrelevant to AI-based education. For instance, “The development of artificial intelligence has entered a new stage.” [30]. Finally, the selected policy statements are manually annotated under the coding framework (see Table 2). For policy content that relates to multiple analytical dimensions, the primary code is determined according to its core policy function to ensure the standardization and consistency of coding. In spite of a degree of arbitrariness involved in the decision-making for the codification, both authors complete the coding independently of each other and agree on over 97% of cases. For those discrepancies, we have an in-person discussion until a consensus is achieved. Furthermore, this paper provides representative coding samples in the Appendix A to enhance the transparency and verifiability of the research process.

To avoid overinterpreting any single document, this study focuses on coding types that recur across multiple policy documents and are highly relevant to higher education. On that basis, national policies are extracted into core characteristics in line with the analytical coding framework. The following sections conduct a cross-national comparison of the differentiated educational policy orientations of the three countries in the field of AI and further generalize explanatory policy patterns. Taking these steps, the study attempts to secure contextual integrity and strengthen the process standardization for an increased methodological reliability of textual analysis.

4. Policy Analysis

4.1. Policies of China

Since their inception in 2017, China’s AI policies have evolved from top-level design to practical implementation in educational settings and then down to institutionalized and long-term deployment. The initiative of such kind has profoundly shaped the digital intelligence transition, as indicative of the integration of higher education with AI in

Table 3. The evolution of China’s artificial intelligence policies.

Stage	Policy Document	Policy Foci	Core Characteristics
Elementary Planning (2017)	Next Generation Artificial Intelligence Development Plan (State Council) [30]	Top-level design and systemic layout of the national AI strategy	Establishment of a national AI strategic system and a “three-step” national strategy framework
Policy Promotion (2018)	Action Plan for Artificial Intelligence Innovation in Higher Education Institutions (Ministry of Education) [31]	Construction of AI disciplinary system and science and technology innovation system in higher education	Coordinated advancement of discipline distribution and research platforms
	Education Informatization 2.0 Action Plan (Ministry of Education) [32]	Advancement of Education Informatization 2.0 and digital transformation of the education system	Construction of an intelligent education system and a digital governance framework
Domain Deepening (2025)	Outline of the Plan for Building a Powerful Nation through Education (2024–2035) (State Council) [33]	Construction of an education powerhouse and promotion of the education system modernization	Promotion of comprehensive coordination and synergistic, sustained development of AI education across education, science and technology, and talent with the goal of building an education powerhouse
Sustainable Development (2025)	Opinions on Deepening the Implementation of the “AI Plus” Initiative (State Council) [34]	Promotion of the deep integration of AI with all sectors of the economy and society	Focus shift from internal construction of the education system to external societal empowerment, with the emphasis on the pervasive application and technological empowerment of AI across industries

.

Positioning of higher education in the national strategy framework. China’s AI policy texts demonstrate that the national priority is given to affordable AI-driven education. Next Generation Artificial Intelligence Development Plan [30] in 2017 established AI as central to industrial upgrading and economic structural transformation, providing strategic direction for education policy at the macro level. In this context, the functions of education expanded beyond mere talent cultivation to an important institutional support for the national innovation system and modernization. Subsequently, Action Plan for Artificial Intelligence Innovation in Higher Education Institutions [31] explicitly incorporated higher education institutions into the national AI innovation system, making AI education a critical component of national strategy through optimized disciplinary arrangements and adjustments to talent-training mechanisms. Based on this, AI policy ceased to be solely a technical-sector matter and became embedded as an institutional element of the national strategic development framework in a broader sense.

Progressive structure of the policy framework. From the perspective of policy evolution, China’s early AI policies were dominated by overarching strategic plans that clarified trending directions and overall objectives. The subsequent policies set out to highlight the construction of standard systems and regulatory frameworks in evidence of their further implementation through application-oriented documents and demonstration projects. In addition, Education Informatization 2.0 Action Plan [32] promoted “Internet + Education” as the digital infrastructure and institutional conditions to integrate AI technologies into the education system. In the 2024–2025 stage, both Outline of the Plan for Building a Powerful Nation through Education (2024–2035) [33] and Opinions on Deepening the Implementation of the “AI Plus” Initiative [34] further incorporated AI into the long-term agenda for education modernization and national governance. The foci in these policies have progressively shifted from strategic initiatives toward institutional integration and system deepening, thereby forming a relatively clear sequential structure.

Institutionalization of the role of higher education. As for policy instruments, China widely employs a variety of tools, including planning documents, guiding opinions, pilot projects, and demonstration initiatives, to navigate higher education institutions in adjusting their disciplinary structures, curricula, and talent cultivation models. Influenced by such adjustments, higher education institutions act both as implementers of policy and as integral components of the institutional system. Notably, Outline of the Plan for Building a Powerful Nation through Education (2024–2035) [33] proposes establishing a mechanism for adjusting disciplinary configurations aligned with national strategic needs and strengthening the development of emerging and interdisciplinary fields, thereby reinforcing, at the institutional level, functions higher education performs in talent cultivation and research support in the field of AI. Similarly, Opinions on Deepening the Implementation of the “AI Plus” Initiative [34] emphasizes improving the layout of disciplines as well as programs and advancing mechanisms for industry-education integration so that higher education assumes a more defined institutional role in building an intelligent economy and intelligent society. Consequently, higher education is not only a policy implementer but also a key node in the implementation of national strategies.

Technological development and ethical norms in parallel. Alongside institutionalized positive measures regarding the role of higher education, China’s AI policies foreground ethics and social responsibility. These policies have progressively incorporated risk prevention and normative construction into institutional frameworks, while prioritizing technological innovation and educational reform. At the early stage, China’s AI plans highlighted the construction of legal and ethical systems. On the other hand, subsequent national documents also called for strengthening data security and algorithm governance. For instance, the Outline of the Plan for Building a Powerful Nation through Education (2024–2035) emphasizes the need to “strengthen cybersecurity safeguards, and reinforce data security, as well as the security and ethics of artificial intelligence algorithms” [33]. Similarly, the Education Informatization 2.0 Action Plan calls to “enhance the cultivation of students’ information literacy and strengthen the cultivation of students’ integrated in-class and out-of-class knowledge, skills, and application capabilities in information technology, as well as their information awareness and ethics” [32]. This dual trajectory of development and regulation reflects an approach that gives weight to social responsibility and institutional stability, when expanding AI education, and embodies a balance between technological advancement and governance constraints.

Global competitiveness of artificial intelligence education. China’s AI policies feature international cooperation and open exchange. As suggested, Action Plan for Artificial Intelligence Innovation in Higher Education Institutions [31] proposes strengthening mechanisms for international talent cultivation and collaboration, and Opinions on Deepening the Implementation of the “AI Plus” Initiative [34] stresses high-level openness and international participation. As such, AI not only serves domestic higher education, industry, and technological upgrading, but also gets incorporated into national strategies aimed at expanding international scientific and technological outreach and participating in global governance. Specifically the role of higher educational institutions within the national global competitive landscape has gradually become more prominent through AI-related talent mobility and the sustainable collaborative networks.

4.2. Policies of Japan

AI policies of Japan originated with the 2017 release of the Artificial Intelligence Technology Strategy [35] with reference to

Table 4. The evolution of Japan’s artificial intelligence policies.

Stage	Policy Document	Policy Foci	Core Characteristics
Elementary Planning (2017)	Artificial Intelligence Technology Strategy (Cabinet Office) [35]	Establishment of the national strategic positioning and top-level design	Strategic positioning, technology drive and interdisciplinary integration insights
Policy Promotion (2019)	AI Strategy 2019 (Cabinet Office) [37]	Quantified target setting, path clarification and institutional implementation reinforcement	Specific action plan and path from general education to professional integration
Domain Deepening (2023–2025)	Policy Package regarding Education and Human Resource Development toward the Realization of Society 5.0 (Cabinet Office) [38]	Deep integration of AI education into the “people-oriented” future social transformation	Industry-academia-research collaboration, education and industry integration, and social problem solving
	Basic Plan for the Promotion of Education (MEXT) [39]	Comprehensive digital transformation and capacity reconstruction of the education system	AI as the core driving force for digital education, systematical infrastructure planning, and teaching models and teacher capacity upgrading
Sustainable Development (2025)	Artificial Intelligence Basic Plan: “Japan Rebooted” through “Trustworthy AI” (Cabinet Office) [36]	Trustworthy AI development and ethical framework construction	Ethical protection, social responsibility, and technological transparency

. The policy document was formulated by the Strategic Council for AI Technology under the Cabinet Office with open access on the website of Council Secretariat. Although not issued as a formal cabinet decision, it nonetheless constitutes a central government-level policy strategy text. The policy-making mechanism established by the strategic council reflects Japan’s institutional approach of cross-departmental coordination and incremental advancement in the field of AI. Subsequently, related policies gradually extend into education and talent cultivation, leading to a basic plan framework with “Trustworthy AI” as its core in 2025 [36].

Close linkage between strategic positioning and societal needs. A synthesis of five core Japanese policy documents demonstrates that the Japanese government, at an early stage, explicitly positioned AI as a foundational technology underpinning national economic functioning, social governance, and long-term competitiveness. In the elementary planning stage in 2017, AI was situated within the broader social transformation of the Fourth Industrial Revolution and associated with strengthening industrial competitive strength. Against this backdrop, AI Strategy 2019 [37] further linked AI with the realization of Society 5.0 and the overcoming of the issues facing Japanese society, thereby making AI an important policy instrument for addressing industrial competition and social structural challenges. Thus, Japan’s early AI policies did not confine AI to a narrow innovation agenda but defined it as a core strategic domain for national structural transformation and the reconstruction of long-term competitiveness.

Generalization and interdisciplinary development in higher education. Japanese policies at higher education level break down silos in the generalization and interdisciplinary integration of AI education. Beyond the traditional fields of engineering or information technology, AI has been introduced into a broader disciplinary spectrum. In a sustainable view, Japanese policies encourage higher educational institutions to incorporate AI-related content across different programs to enhance students’ foundational competence of understanding.

As the policies advances their implementation, the institutional status of undergraduate and graduate schools within the AI talent cultivation system has become increasingly prominent, positioning them as critical nodes that align basic education with high-end research and development systems. Policy documents repeatedly stress the mindset of “AI for Everyone” in basic AI education, which are not specialized and do not have clear boundaries between arts and sciences. They explicitly require the systematic dissemination of mathematical, data science, and AI literacies at the higher education stage. As obviously indicated by Japan’s core assessment of AI education, in an increasingly intelligent society, AI is no longer the professional tool of a restricted technical cohort, but a fundamental competency all citizens need to comprehend social operations and participate in public and industrial activities. By popularizing basic AI knowledge in higher education, the policies double efforts to reduce the uncertainty generated by technological change.

Centralized guidance with limited institutional flexibility. In the initial stage (2017–2019), the Japanese government delineated the overarching direction and incorporated AI education into the national medium- and long-term sustainable development plans. Some concrete policy instruments were employed, such as formulation of curriculum standards, nationwide dissemination of textbooks, and an accreditation system for educational programs. To this end, the implementation of AI education was promoted within a unified framework across higher education institutions. At the stage of institutional promotion, the policy foci shift toward how the education system can realize transformation within existing structures. Particularly, Society 5.0 [38] calls for a shift in the system of education and talent cultivation and thus advances reforms from an all-Japan and cross-departmental perspective. This institutional configuration both strengthens strategic guidance at the national level and preserves the flexibility of higher education in defining specific implementation pathways. More importantly, AI-driven education reform is expected to remain stable alignment with national development goals, adaptive institutional evolution in response to technological and industrial changes, and long-term sustainable capacity-building for intelligent education, provided that the overarching objectives of education remain converging. This institutional configuration, to some extent, also corroborates the observation in prior research that value declarations do not necessarily translate fully into concrete institutional arrangements [18]. It demonstrates that when national-level normative principles descend to the institutional level of higher education, their implementation is constrained by factors such as organizational capacity, resource allocation, and path dependency, resulting in a pattern of “selective implementation.”

Value orientation and risk awareness. While encouraging technological innovation, Japan’s AI policies have consistently put high on the agenda the parallel advancement of human-centered values and risk governance. As indicated in AI Strategy 2019 [37], “Human-Centric AI” is defined as a fundamental principle to incorporate AI development within a framework that respects individual dignity and the public interest. Artificial Intelligence Basic Plan [36] further advocates the construction of “Trustworthy AI”. Essentially it is stressed to foster innovation, elevate risk management, and reinforce institutional safeguards at the same time. Also worth highlighting are values of such an orientation not only at the level of macro governance but also deeply in educational policy. Thus, AI education is regarded as an important means to enhance social collective judgment and risk awareness, rather than merely a tool to cultivate technical skills. Additionally, AI development is required to proceed within frameworks of transparency, fairness, and social acceptability, thereby establishing a relatively stable balance between technological advancement and social norms. However, although some higher educational institutes have introduced documents such as guidelines for the use of AI and have incorporated content related to AI ethics and data responsibility into their curricula, the overall educational focus remains predominantly on cultivating technical competencies. Modules in humanities and ethics are often appended in a supplementary manner, which has resulted in the principle of “human-Centric” not being fully translated into actionable institutional arrangements in practice.

Lifelong learning and societal extension. Japan’s AI policies highlight the importance of expanding AI education across all levels of society and enhancing the overall AI literacy of the Japanese population through a lifelong learning system. In particular, Basic Plan for the Promotion of Education [39] stresses the pivotal strategies of constructing an environment where people could learn and play active roles throughout life and of promoting the continuous enhancement of digital and innovation capacities across different age groups. Within this framework, AI education serves not only talent cultivation in higher education but also improvement of overall social adaptability as a key mechanism. By strengthening re-education and continuous learning, Japan’s AI-related policies, on the one hand, attempt to reinforce the long-term capacity reserves of social members during the transition to an intelligent society. On the other hand, AI education will be in play to extend from schools into the broader society, enabling all citizens to adapt to a rapidly developing digitalized and intelligent society.

4.3. Policies of South Korea

South Korea’s national strategic framework for AI was established in 2019. South Korea has progressively constructed a development trajectory characterized by industry-driven growth and strengthened governance through continued policy implementation, intensified digital transformation, and long-term planning. This trajectory has imposed structural impacts on the integration of higher education and AI, as seen in

Table 5. The evolution of South Korea’s artificial intelligence policies.

Stage	Policy Document	Policy Foci	Core Characteristics
Elementary Planning (2019)	National Strategy for Artificial Intelligence (MSIT) [40]	Establishment of the national AI strategy framework	Strategic awakening; formulation of a macro-level vision; whole-chain systematic planning
Domain Deepening (2022–2023)	Opening up “the Era of Tailored Lifelong Learning Enjoyed by All” (Ministry of Education) [41]	Framework for a lifelong learning society and capability renewal	Reconstruction of societal capabilities; differentiated learning; AI-supported lifelong learning
	Digital-driven Education Reform Plan (MSIT) [42]	Digitalization and intelligent transformation of the education system	Reconfiguration of teaching models; transformation of teacher roles; design of highly operational projects
Policy Promotion (2024)	National AI Strategy Policy Directions (MSIT) [43]	Concentrated advancement of national quantitative targets	Quantitative target setting; public–private collaboration; comprehensive mobilization
Sustainable Development (2024)	Strategy to Realize Trustworthy Artificial Intelligence (MSIT) [44]	Construction of a trustworthy AI governance system	Synchronous governance; ethics embedment; forward-looking risk management

.

Strategic layout in global artificial intelligence competition. In South Korea’s overall strategy, AI is designated as the core lever to achieve a leap in its national status and regain global economic and geopolitical influence. To that end, National Strategy for Artificial Intelligence [40] accentuates that the national overall competitiveness should be enhanced through technological research and development, industrial application, and talent cultivation system. At this stage, policy priorities are given to consolidate the institutional foundations and industrial ecosystem for AI development, thereby laying a framework for subsequent strategic upgrading. Moreover, National AI Strategy Policy Directions [43] further reinforced a global-competition orientation. It is worth setting objectives, such as “Emerging as one of the top three AI powerhouses to become a Global Pivotal State” [43], and accelerating the scaled deployment of AI through expansion of AI computing infrastructure, national flagship project layout, and promotion of cross-sector “AI + X” applications. Compared with the elementary planning stage, this stage shifting from development planning to competition intensification stands out owing to the international competitive positioning and the enhanced national comprehensive strength.

Deep integration of educational reform and artificial intelligence technology. The policies of South Korea exhibit a distinct trend of technology-embedded reform in the education sector. Digital-driven Education Reform Plan [42] indicates “AI-driven digital textbooks” and “personalized learning” as core directions, literally embedding AI technologies within classroom instructional structures. Under this proposal, AI functions not only as curricular learning content but also as a pedagogical tool for learning-data analysis, progress management, and personalized feedback, promoting a shift from unified instruction to differentiated learning modes. The reform places a high value of reconstructing the teaching process through digital technology, by which AI is rendered part of the education operational mechanism rather than merely a specialized domain of knowledge. Teachers’ roles are correspondingly adjusted from traditional knowledge transmitters to learning facilitators and data supporters. Consequently, AI education in South Korea demonstrates a fusion characteristic that aligns technology with industrial development.

Regulation and ethical safeguards for high-risk technologies. Strategy to Realize Trustworthy Artificial Intelligence [44] advocates a “technology, system, ethics” triad as the three pillars for constructing a “trustworthy AI” framework. Thus, in tandem with AI-driven, innovation technological growth should get grounded in institutional and ethical frameworks to strengthen capabilities for risk identification and governance. In this view, AI in the educational context is no longer driven solely by efficiency and scale but is advanced within a clearly defined regulatory structure. AI ethics and safety are likewise key components of South Korea’s AI policies. By establishing AI safety agencies and enacting medium- to long-term legislation to protect user rights, the policies provide institutional safeguards for the sustained and healthy development of AI. The policy documents unfold South Korea’s responsible stance toward social risks and ethics alongside the rapid application of AI technologies. Similar to Japan, South Korea’s “trustworthy AI” framework reflects a high level of importance placed on risk governance and institutional norms at the policy level. However, the specific implementation pathways within its higher education system remain to be further clarified, revealing a potential gap between the ethical framework and its institutional operationalization.

Government leadership and cross-departmental coordination mechanisms. In promoting AI education policy, South Korea draws attention to governmental leadership and cross-departmental coordination. National AI Strategy Policy Directions [43] proposes the significance of establishing a “National Artificial Intelligence Committee” to strengthen the coordinated and strategic implementation of the national policy, enabling AI to thrive within a unified framework. In this mechanism, the cultivation of AI talents is incorporated as a priority of the national AI strategy and is planned to interact with industrial innovation and technological development. At the same time, the policy stresses the strategies of building the AI ecosystem through public–private collaboration, and creating a cooperative structure among government, industry, and research institutions for talent growth and technology application. By enhancing the training and capacity-building of AI professionals, the linkage between the education system and industry demand is institutionally reinforced. Consequently, South Korea’s AI education presents a model of advancement centered on national leadership and supported by multi-stakeholder collaboration.

Educational equity and social inclusion. South Korea’s AI policy highlights social inclusion and universal access to learning opportunities in the education sector. In 2022 the concept was put forward in relation to “Opening up ‘the era of tailored lifelong learning enjoyed by all’” [41], defining lifelong learning as an institutional arrangement for all citizens rather than supplementary education privileged to specific groups. Within this conceptual framework, AI and digital technologies are regarded as key instruments to enhance accessibility to learning and provide personalized support. Regardless of ages and occupational backgrounds, all citizens can participate in the digital transformation process through re-skilling and up-skilling mechanisms, thereby narrowing skill gaps and structural inequalities. It is underscored to link lifelong learning with workforce capability enhancement so that AI education, while supporting individual capacity enhancement, advances in coordination with industrial structural adjustments. Not unexpectedly, AI education is not only oriented toward training high-end talent but also incorporated into institutional arrangements aimed at improving overall societal adaptability.

5. Comparing the Sustainable Role in the AI-Related Policies of Higher Education

The previous section systematically presents the core characteristics of AI-related policy texts from China, Japan, and South Korea based on a unified coding framework. Building upon such comparative analysis, this section systematically addresses the research questions and synthesizes and substantiates the main analytical conclusions through the findings of the cross-national comparison. The following eight comparative dimensions, as illustrated in

Table 6. Comparison of AI-related policy characteristics of China, Japan, and South Korea.

Dimensions	China	Japan	South Korea
National Strategy Embedment	National core strategic technology; service to overall development and governance system	Foundational support for intelligent society operation	Key lever for global competitiveness and national advancement
International Participation Model	Two-way talent mobility and global governance participation	Export of social models and value concepts	Drive of Technology catch-up and industrial cooperation
Policy Advancement Pathway	Anchoring in national strategy → institutional embedment → trinity integration → social empowerment and institutional export	Social demand drive → standardization → social vision integration → trust building and value leading	Global competition orientation → technological integration advancement → policy acceleration → industry–academia collaboration
Role of Higher Education	Institutionalized execution entity in the national innovation system	Core mechanism for enhancing society’s overall understanding and adaptability	Key platform for technology deployment and talent supply
Higher Education Reform Mechanism	Parallel advancement of disciplinary organization and resource allocation	Curriculum standard guide and autonomous implementation by higher education	Direct embedment of technical solutions into the educational process
Talent Cultivation Orientation	Interdisciplinary talents in “Artificial Intelligence + X”	General education and basic literacy for all population	Personalized, technology-enabled learning capacities
Governance and Development	Priority on development objectives; parallel layout of ethics and law; emphasis on long-term institutional stability	Human-centered value guidance and preventive governance; emphasis on socially sustainable operation	Systematic risk management under a trustworthy AI framework; technology application enhancement
Policy Extension to Society	Expansion from higher education to industry and society; service to long-term national development strategy	Higher education → lifelong learning → whole society coverage; support for sustainable adaptation of social structure	Coordination among education, fairness, and labor-force transformation; reinforcement of social inclusion and capability renewal

, elucidate the theme-based features of the policies across the three countries. These dimensions reveal deeper divergences in educational governance orientations and strategic logic.

Our analysis of the policy texts reveals that certain dimensions are intrinsically linked in terms of logic. For instance, Policy Positioning, together with International Participation, reflects national strategy embedment. Policy Advancement Pathway and Realization Mechanism embody the organizational logic of policies. Role of Higher Education and Talent Cultivation Orientation both constitute differences at the level of educational philosophy. Governance and Development and Policy Extension to Society collectively capture each country’s understanding of the relationship between development and regulation. Based on these intrinsic logical associations, the following discussion clusters the eight comparative dimensions into four aspects at a relatively macro level, namely strategic positioning, policy architecture, educational philosophy, and governance model. Simultaneously, a cross-analysis of relevant policy characteristics is conducted by integrating the three analytical perspectives: governance challenges of GenAI, transformation of competency structures, and ethics and governance.

5.1. Strategic Positioning

All three countries elevate AI education to the level of national strategy, but they differ in emphasis and strategic orientation. To begin with, China is regarded as a strategy-driven nation. AI is explicitly positioned as a core strategic technology for national development, with emphasis on its strategic support for national competitiveness and economic transformation. Second, Japan is typical of a value-led nation. Its approach reflects a mindset of building social infrastructure. Also, its education policies focus more on enhancing citizens’ overall understanding of and adaptability to AI, aiming to lay a broad human and cognitive foundation for an intelligent society and to better integrate AI into social life while raising public literacy. Different from the previous two countries, South Korea is more oriented toward technological acceleration. Its traits unveil a model driven by global competition, aiming to build South Korea into a global pivotal state as one of the top three AI powerhouses. Prominently South Korea relates its education to industrial development goals, which leads to a focus on enhancing global competitiveness through technological innovation and education reform. From a sustainability perspective, the educational systems of all three countries are embedded in long-term national development structures. Undoubtedly, their functions are no longer limited to knowledge transmission but serve as institutional foundations supporting economic structural adjustment and social transformation. This trend aligns with one thread of research that views higher education as assuming structural public functions within a sustainable development framework [45]. Such a strategic integration provides policy assurance for stable investment in and long-term development of higher education.

Moreover, all three countries recognize AI as a domain of global competition, but they differ in their strategies for participating in globalization. Underlining institutional participation, China encourages students to study abroad and attracts foreign students to promote two-way talent flows. Also, China actively engages in global governance issues, such as international standards and ethical norms for AI, striving to exert influence in the formulation of international rules. In a different vein, Japan’s globalization approach places greater emphasis on the diffusion of social models. Advocating a human-centered development philosophy, Japan regards education as a key path for exporting AI-related social values and governance experience. By contrast, South Korea demonstrates an open strategy driven by technological catch-up and industrial cooperation. As regards its global participation, South Korea closely relates the national industrial objectives to the rapid integration into global industrial chains and innovation networks. Leveraging the partnerships between public and private sectors, the establishment of international innovation clusters, and the commercialization of domestic technologies, South Korea realizes its strategic goal of becoming a global AI powerhouse.

All in all, the differences among the three countries largely showcase variations in institutional definitions of the educational function under distinct developmental logics, rather than those in the choice of policy instruments. The strategic embedment of AI in higher education could be understood as an institutional continuation and adjustment of the national developmental trajectories.

5.2. Policy Architecture

All three countries have established policy architectures spanning from macro-level strategies to concrete implementations, but obvious discrepancies exist in their logics and emphases. These policy architectures, to a certain extent, reflect the institutionalized pathways that various countries have adopted in addressing the governance challenges of AI. As for China, its policy architecture is highly systematized and incremental. Typically its policy instruments include planning documents, discipline development, resource allocation, and pilot projects. In this way, China foregrounds the institutional embedment of AI education into the national innovation system through formal mechanisms. The Chinese approach is characterized by strong implementation capacity and concentrated strategic direction. Related studies indicate that governance of Chinese higher education has consistently exhibited a strong national orientation, and provided continuous guidance and coordination of the higher education system through policies and fiscal instruments [46]. This incremental institutional construction demonstrates substantial capacity for institutional integration. Similarly, Japan’s policy architecture underscores top-level design and social diffusion at the same time. For instance, basic plans are formulated by specialized councils to set overall direction, while instruments, such as curriculum standards, textbook promotion, and program accreditation, guide nationwide implementation. The features of the Japanese approach include balancing national guidance with higher education autonomy and placing importance on lifelong learning, thereby extending AI beyond higher education into the whole society. By comparison, South Korea’s policy architecture focuses on technological integration and rapid deployment. It is worth noting the deep integration of AI technologies into education reform processes, which is representative of the development of personalized learning systems. South Korea’s policies reflect a notable technology-driven pathway for education reform and accelerate the transferring of technical solutions into educational practice through government direction and collaboration between public and private sectors. Research shows that cross-national differences in the design and implementation approaches of AI education policies are closely related to institutional inertia and policy priorities of each country [9]. Generally speaking, the architectural complexity of AI education policies has become a key dimension for assessing a nation’s capacity of institutional integration and the long-term stability of its strategic orientation in the context of technological change.

5.3. Educational Philosophy

Of all the AI-related policy documents explored, the cultivation models display distinct requirements for educating AI talents in the three countries. China stresses interdisciplinary and composite disciplinary capabilities, vigorously promoting an “AI + X” model for training multidisciplinary talents. Accordingly, AI education is understood as cultivating a general-purpose technological competence that must be deeply embedded across varied domains, combining high specialization with broad interdisciplinarity. By comparison, Japan advocates general education and the dissemination of foundational competencies. For one thing, Japan aims to equip the entire undergraduate population with basic AI literacy. For another, Japan heightens the requirements to cultivate modern citizens with fundamental mathematical, data science, and ethical literacies to meet the widespread needs of a sustainably developing intelligent society. By popularizing basic AI literacy, Japan seeks to build a long-term capacity base at the societal level and reduce the structural disruptions caused by technological replacement. This approach reflects its emphasis on the capacity of education system for sustained adaptation. South Korea, however, focuses on personalized and technology-enabled learning. By tailoring instruction to students’ learning in all settings, South Korea leverages AI technologies to transform educational models and provide students with individualized learning approaches and resources. Its training objectives highlight individual innovation, collaboration, and problem-solving skills within AI-embedded environments, where technology functions both as subject matter and as a means to reshape the educational process. The personalized learning and lifelong learning under the policy frameworks have been established to help enhance the capacity of workforce for continuous updating amid rapid technological iteration and strengthen long-term alignment between education and employment systems. From the perspective of competency structure transformation, the differing connotations of educational philosophies among the three countries also reflect variations in the adjustment of talent cultivation objectives and competency structures within the context of AI.

Note that by comparison, educational discourse of the policy documents is rarely a purely technical articulation rather than carry implicit value premises. Policy debates on AI often construct differentiated educational objectives and institutional expectations through divergent definitions of “capability,” “innovation,” and “responsibility” [47]. In this sense, differences in the three countries’ educational logics are manifested not only in the differing prioritization of development objectives, but also in their distinct understandings of the social significance of AI.

5.4. Governance Model

Policy documents from the three countries all recognize that despite great opportunities for technological innovation and development, AI entails potential risks and governance challenges. China adopts an institution-first strategy to advance AI, employing the national strategy to navigate AI development. In the higher education, measures, such as establishing AI as a first-level discipline and promoting interdisciplinary “AI + X” training, have been adopted to systematically embed ethical awareness and legal literacy in talent cultivation. In contrast, Japan places greater emphasis on embedding value orientation and normative awareness. Specifically, through the “Human-Centric AI” principle and general education reforms, AI ethics and social responsibility are integrated into student competency. It is therefore acknowledged that technical innovation capabilities and value-judgement skills mature concurrently. In a similar vein, South Korea follows an approach of systemic assurance on the national part. Institutionalized measures, as indicated in Strategy to Realize Trustworthy Artificial Intelligence [44], contribute to promoting AI digital textbooks and teacher digital literacy training, while incorporating ethical norms and data security awareness throughout the talent development process. Although the policies in South Korea accelerate technology deployment, they rely on higher education to internalize governance concepts as professional competencies in future personnel. In this regard, such a model advances innovation acceleration and risk management in parallel. In turn, the model forges a synergy between guarantees reinforced by the national system and efforts made by educational institutions.

Through the lens of ethics and governance, the three countries exhibit varied pathways between value orientation and institutional norms. All of them, in advancing their strategies, concurrently give prominence to risk governance and ethical norms. Their focus is not set on short-term regulation but the incorporation of AI into sustainable governance architectures under the institutionalized frameworks to preserve the long-term stability of educational systems and social structures. As previous research on AI governance indicates [14,48], only when embedded within concrete institutional arrangements and policy mechanisms will ethical principles exert practical constraint and guidance on technological development. Likewise, risk governance is typically institutionalized through normative frameworks and mechanisms for allocating responsibility, instead of remaining at the level of conceptual principles [49,50]. Against this backdrop, the differences in higher education institutional arrangements are articulated, to a large extent, in the national AI strategies of the three countries. In the context of technological change, the approaches of governance models, in fact, are divergent for AI development in a long run.

5.5. Moving Toward Sustainable Higher Education Policies for Artificial Intelligence

The preceding systematic comparison of AI policies in China, Japan, and South Korea revealed three distinct policy paradigms. These paradigms not only reflect pluralistic choices of each country in strategic positioning, policy architecture, educational philosophy, and governance model, but also unpack different explorations into sustainable development approaches for higher education in the AI era. As United Nations Sustainable Development Goal 4 (SDG4) proposes “ensure inclusive and equitable quality education and promote lifelong learning opportunities for all,” [51], AI education for sustainable development pertains not only to the enhancement of technical capabilities, but also to the harmonization of institutional guarantees with social equity objectives [52], aligning with the pedagogical design level with sustainable development frameworks [53]. The two components of national commitments are inextricably intertwined in the following three dimensions: institutional sustainability, social sustainability, and innovation sustainability.

At the institutional level, a stable institutional environment plays a decisive role in promoting AI in higher education. The incorporation of AI education into the national innovation system through strategic planning and resource allocation helps establish sustained policy expectations and structural support, thereby ensuring continuity in talent cultivation and knowledge accumulation.

At the societal level, AI education is required to get embedded within the national innovation framework by means of policy planning and resource allocation. Such synergistic action contributes to consistent policy signaling and structural underpinnings. Therefore, it safeguards the coherence of workforce development and knowledge building. In this process, higher education undertakes the task of raising professional competencies and engages in reshaping societal interpretations and normative expectations of technological change.

At the innovation level, given the rapid pace of technological iteration, the educational system must retain an appropriate capacity for sustained adaptation. In other words, the responsiveness of the educational system should be boosted as a result of promoting the renewal of instructional models empowered by AI-driven technologies. Yet in tandem with innovation efficiency unlocked by the responsiveness, serious concerns that deserve ample attention are the risks arising from technological dependency and imbalanced resource distribution.

To sum up, the above analysis reveals the sustainable role of higher education in national AI strategies from three dimensions: institutional, social, and innovative, which manifests as the synergistic interaction of institutional support, the embeddedness of value orientation, and the mechanism for innovation drive. Differences exist in higher education system arrangements in relation to the national AI strategies of China, Japan, and South Korea. Arguably their choices regarding long-term development approaches are distinct in the context of technological transformation. Also their policy models influence the competitive landscape of the AI industry and profoundly reshape the structural role of higher education systems in future societies. Most importantly, their commitments are of value in response to SDG 4.

6. Concluding Remarks

6.1. Findings and Pedagogical Implications

Set against the context of advancing higher education reform in the AI era, this study has sought to contribute to revealing the policy logics and approaches adopted by China, Japan, and South Korea. Stemming from differing emphases in their national AI development objectives, the major divergences reflect deeper structural influences arising from political systems, educational traditions, and socio-cultural contexts. In response, our findings based the comparative analysis on policy documents suggest that distinct educational ecosystems of the three countries are configured under their respective AI policies. Now more than ever, China boosts its capacity for strong mobilization and institutionalization, while Japan upholds its tradition of building social consensus on and reflecting value of national policies. What sets South Korea apart is its priorities for cutting-edge technologies and implementation efficiency. On the other hand, the three nations are all ramping up efforts to strike a balance among technological empowerment, ethical constraints, educational innovation, social integration, and sustainable development. Understanding these differences and commonalities is of practical significance for other countries to build sustainable education policy frameworks in the AI era, cultivate future talent capable of driving long-term societal progress, and participate in global AI governance. Notably, the policy analysis reveals a more pronounced priority and institutionalization of AI in national policies compared to other emerging fields like educational neuroscience. This disparity suggests that the developmental pathways of different technological domains within higher education may be significantly shaped by how they are embedded within national policy frameworks.

Furthermore, from educational practice, the analysis in this study also offers some insights. The differences in AI higher education policies among China, Japan, and South Korea are not only reflected in institutional arrangements and governance models but also, to a certain extent, influence teaching reform pathways, classroom practice, and student competency cultivation. For instance, policy designs centered on competency orientation drive higher education to place greater emphasis on integrating interdisciplinary skills and AI literacy in curriculum design and teaching methods, while those on governance and regulation set clearer boundaries and ethical requirements for the use of AI tools in the classroom. From this perspective, the “sustainable role” of higher education in AI development is manifested not merely in its supportive function at the institutional level but in its capacity to continuously adapt to technological changes and reshape talent cultivation models through teaching practices. This further indicates that policy orientations at the national level can permeate the education system, influencing classrooms and learning processes, thereby exerting a profound impact on the long-term development of higher education.

6.2. Limitations and Further Research

We coded, compared and analyzed representative policy documents in China, Japan and South Korea. As with any study, however, our study has some limitations in interpreting governance logic and strategic intentions embedded in the texts. First, the analysis is based solely on national policy documents and does not capture local practices, institutional responses, or the effectiveness of policy implementation. Second, rapidly evolving AI environment necessitates further monitoring of dynamic adjustments to policies in the three countries, even other countries outside Asia. While text analysis can elucidate features of policy discourse, it is limited in fully capturing the complex interactions and unintended interplay that arise during policy implementation. Finally, regarding methodology, although a systematic coding framework is employed, qualitative analysis remains inevitably subject to the influence of the researchers’ subjective judgment. Future research could integrate quantitative analytical methods, expand the scope of policy samples, incorporate more countries or regions for comparative analysis, and place greater emphasis on the actual effects at the level of policy implementation, thereby bridging the gap between policy texts and practical outcomes.

To scale sustainability of AI globally in educational context, future research is encouraged to conduct comparative case studies of policy implementation and examine the gaps between policy objectives and actual outcomes. On the other hand, future research is expected to generate a more nuanced view of other important regions, such as Southeast Asia, Europe and North America. An enlarged scope of investigation will more comprehensively trace the evolution of the roles played by AI in influencing policy-making at national or local levels for the sake of a sustainable accessibility to technological resources of higher education.