Sustainable AI governance in Islamic higher education: a maqāṣid-based framework aligning global standards with Saudi vision 2030

Abstract

Higher education is undergoing rapid transformation driven by artificial intelligence (AI), creating significant opportunities for sustainable development alongside complex governance and ethical challenges. Existing scholarship on AI in higher education can be broadly categorized into SDG-aligned frameworks, principle-based governance models, and value-based approaches; however, these strands remain insufficiently integrated at the institutional level and lack operational clarity across the AI lifecycle, particularly in procurement and post-deployment audit stages. This study examines how maqāṣid al-sharīʿah can be systematically operationalized as a framework for sustainable AI governance in Islamic higher education, and how such an approach supports socially sustainable and ethically aligned digital transformation under Saudi Vision 2030. The study adopts a systematic mapping approach with integrative synthesis, analyzing 143 peer-reviewed studies retrieved from Scopus, Web of Science, and ERIC (2015–2026). A combined deductive–inductive coding framework identifies structural gaps between ethical principles and institutional governance practices across four AI lifecycle stages: procurement, design, deployment, and audit. The findings reveal a persistent principle–implementation gap most pronounced in procurement and audit processes. In response, the study develops a maqāṣid-based governance framework that translates core ethical objectives into measurable, sustainability-oriented indicators embedded within lifecycle decision points. The framework is structured as a prototype institutional scorecard enabling university administrators and policymakers to evaluate AI adoption within risk management and decision-making workflows. The proposed model advances social sustainability by promoting accountability, equity, and culturally coherent AI integration, offering a structured pathway for aligning institutional AI practices with the strategic priorities of Saudi Vision 2030.

1 Introduction

Higher education is undergoing rapid transformation driven by advances in digital technologies and artificial intelligence (AI) (Chen et al., 2020), including learning analytics and generative systems that are reshaping teaching, assessment, and institutional decision-making (Holmes et al., 2019; Luckin et al., 2016). These developments position AI as both a catalyst for educational innovation and a source of complex ethical, social, and governance challenges that require systematic institutional responses. In parallel, global policy agendas—most notably the United Nations Sustainable Development Goals (SDGs)—emphasize the role of higher education in advancing inclusive and sustainable societies, with AI frequently framed as a strategic enabler of quality education, innovation, and institutional effectiveness (Floridi et al., 2018; United Nations, 2015). Governance, in this context, functions as a foundational pillar of institutional sustainability (Winfield and Jirotka, 2018), ensuring that rapid digital transformation does not compromise ethical integrity, social equity, or long-term institutional resilience.

Figure 1

Despite growing recognition of AI's transformative potential, the literature on AI governance in higher education remains fragmented across three distinct analytical levels: SDG-oriented institutional frameworks, principle-based ethical and governance models, and value-based approaches rooted in religious or philosophical traditions (Hashim and Langgulung, 2008; Jobin et al., 2019; Zawacki-Richter et al., 2019). While each of these streams offers important contributions, they are often developed and applied in isolation, with limited translation of ethical principles into institution-level processes for procurement, deployment oversight, and audit. This fragmentation is particularly evident in the disconnect between high-level normative commitments and practical implementation guidance. As synthesized in recent reviews, operational governance remains uneven across AI lifecycle stages, reflecting a persistent “principle–implementation gap” in higher education AI governance (NIST, 2023; Selwyn, 2019; UNESCO, 2021). While global frameworks define what constitutes ethical AI, they rarely provide actionable pathways for embedding governance mechanisms within institutional decision processes, particularly in procurement and post-deployment audit stages.

Global initiatives such as UNESCO's principle-based guidelines and NIST-informed risk governance approaches offer important normative and technical foundations; however, they provide limited institution-specific guidance for operationalizing ethical criteria across end-to-end AI lifecycle processes in higher education (e.g., procurement thresholds, design constraints, deployment monitoring, and post-implementation auditing) (OECD, 2019; UNESCO, 2021). Conversely, institution-specific models tend to offer contextually grounded procedures but remain limited in transferability beyond their immediate institutional environments (Selwyn, 2019). Consequently, the field lacks integrated governance frameworks that combine normative grounding, operational specificity, and contextual adaptability across institutions with comparable value commitments.

Within this gap, value-based approaches grounded in Islamic intellectual traditions—particularly maqāṣid al-sharīʿah (the higher objectives of Islamic law)—offer a potentially systematic ethical framework for AI governance. Centered on the preservation of religion, life, intellect, lineage, and property, the maqāṣid framework provides a teleological structure oriented toward benefit (maṣlaḥah) and harm prevention (mafsadah). Recent studies have explored conceptual alignments between maqāṣid principles and contemporary AI ethics frameworks and have proposed preliminary mappings to governance models; however, these efforts remain largely theoretical and lack systematic operationalization at the institutional level (Auda, 2008; Dusuki and Abdullah, 2007; Hashim and Langgulung, 2008; Hassan et al., 2022; Kamali, 2008; Madani et al., 2026). This study addresses this conceptual limitation by advancing a structured approach to translating maqāṣid principles into measurable governance indicators applicable to AI systems in higher education contexts.

This gap is not confined to a single intellectual tradition. Both Western and Islamic AI governance models tend to articulate ethical commitments at a normative level without systematically embedding them into institutional processes, thereby reinforcing the need for operationally grounded frameworks.

In this study, “Islamic universities” refers to higher education institutions whose institutional mission, governance mandates, or curricula are explicitly informed by Islamic ethical and epistemic frameworks, regardless of geographic location. This definition emphasizes institutional orientation rather than geography, thereby enabling the framework's applicability across diverse contexts that share similar value-based commitments.

The need for an operationally grounded governance framework is particularly salient in the context of Saudi Arabia, where Vision 2030 articulates an ambitious agenda for digital transformation, human capital development, and educational excellence. The Vision's emphasis on building a “vibrant society” and a “thriving economy” through technological innovation—while preserving cultural identity—creates both opportunities and imperatives for developing culturally grounded approaches to AI governance in higher education (Kingdom of Saudi Arabia, 2016). In practice, however, technological adoption and ethical governance are often treated as parallel rather than integrated processes, limiting their collective contribution to sustainable development.

Against this backdrop, this study addresses the following research question: How can maqāṣid al-sharīʿah be systematically operationalized into measurable governance indicators for sustainable AI deployment in Islamic higher education, and how does such an approach address the principle–implementation gap while supporting the strategic objectives of Saudi Vision 2030?

To address this question, the study employs an integrative literature review using predefined inclusion criteria and thematic content analysis across major academic databases (detailed in the Methods section). This approach enables the synthesis of diverse theoretical, policy, and empirical streams that cannot be adequately captured through strictly systematic review designs. To enhance transparency and analytical rigor, the Results section presents a typology of identified framework families alongside a lifecycle coverage map (procurement–design–deployment–audit), highlighting where operational guidance is concentrated or lacking across the literature.

The study ultimately advances an integrated Maqāṣid al-Sharīʿah-based framework for AI governance, operationalized through a prototype institutional scorecard. The framework is designed to be interoperable with global AI governance standards while translating ethical principles into actionable governance mechanisms. Specifically, it provides (i) lifecycle-based governance checkpoints, (ii) measurable indicators aligned with maqāṣid objectives, and (iii) a structured risk and decision matrix for institutional evaluation. By moving beyond abstract ethical prescriptions toward implementable governance tools, the proposed framework contributes to socially sustainable, ethically coherent, and institutionally resilient AI integration in higher education, in alignment with the strategic priorities of Saudi Vision 2030.

2 Materials and methods

2.1 Research design

This study adopts an integrative literature review methodology to examine how artificial intelligence (AI) governance frameworks in higher education can be operationalized through value-based approaches, with a specific focus on maqāṣid al-sharīʿah. This approach enables the synthesis of conceptual, empirical, and policy-oriented studies to generate new theoretical insights and identify structural gaps in the literature.

The analysis combines thematic content analysis with cross-framework comparative mapping, allowing systematic identification of recurring governance patterns, limitations, and operationalization gaps. The aim of this integrative review is to bridge disparate conceptual, policy, and governance-oriented streams into a unified operational framework for AI governance in higher education.

2.2 Data sources and search strategy

A structured search was conducted across three major academic databases: Scopus, Web of Science, and ERIC. The search covered publications from 2015 to early 2026, including articles published online ahead of print. References dated 2026 in the manuscript reflect studies retrieved as online-ahead-of-print at the time of the search and should not be read as dating errors.

2.3 Inclusion and exclusion criteria

Studies were selected based on predefined criteria.

2.3.1 Inclusion criteria

• Peer-reviewed journal articles (conceptual and empirical)

• Policy frameworks and institutional governance models

• Studies addressing AI or digital technologies in higher education

• Publications in English or Arabic

• Studies engaging with ethics, governance, or sustainability

2.3.2 Grey literature

Selected grey literature—such as institutional reports and intergovernmental policy documents—was included where directly relevant to AI governance in higher education, cited within peer-reviewed scholarship, and attributable to a recognized institutional or intergovernmental source.

2.3.3 Exclusion criteria

• Studies outside higher education contexts

• Purely technical AI studies without governance or ethical dimensions

• Non-analytical opinion pieces

• Grey literature from non-authoritative or non-peer-engaged sources

The analysis was conducted in the original languages (English and Arabic) to preserve semantic and conceptual nuances, particularly in value-based and ethical interpretations.

2.4 Study selection process

Records retrieved from the databases were exported to Zotero for management. Duplicates were removed using a combination of automatic detection and manual verification. The study selection followed a structured multi-stage screening process: initial database searches across Scopus, Web of Science, and ERIC yielded 512 records (Scopus: n = 247; Web of Science: n = 198; ERIC: n = 67). After deduplication, 421 records remained. Title and abstract screening retained 236 studies, with 185 records excluded primarily due to absence of a higher education context or lack of governance and ethical dimensions. Full-text assessment excluded a further 93 studies, predominantly due to lack of relevance to governance or higher education contexts. The final corpus comprised 143 peer-reviewed studies, supplemented by selected policy documents.

2.5 Data analysis

The analysis proceeded in two stages.

2.5.1 Thematic content analysis

Inductive coding complemented this structure by identifying emergent themes from the literature, particularly recurring operational gaps such as procurement deficiencies, audit limitations, and cognitive dependency risks.

A codebook was developed to structure this process, resulting in a cross-matrix used to systematically map governance coverage across the reviewed corpus. Coding was conducted by the author using a structured single-coder design. To enhance analytical consistency and minimize interpretive drift, three verification procedures were applied: (1) iterative cross-checking of coded items against the codebook definitions at regular intervals; (2) periodic blind re-coding of a randomly selected sample (approximately 10% of the corpus) across separate analytical sessions to assess intra-coder consistency; and (3) a structured audit trail documenting coding decisions and boundary cases. The absence of a second coder represents a methodological limitation; inter-rater reliability testing is recommended in future replications of this framework.

2.5.2 Cross-framework comparative mapping

2.6 Framework development approach

2.7 Data availability

All data used in this study are derived from publicly available academic and policy sources, which are fully cited in the manuscript. No new datasets were generated.

2.8 Ethical considerations

This study does not involve human participants or animal subjects and therefore did not require ethical approval.

2.9 Use of generative artificial intelligence

AI-assisted search tools (Elicit) were used to support the systematic database search and initial screening of records. Generative AI tools (Claude, Anthropic) were used for grammar, language editing, and formatting support during manuscript preparation. No AI tools were used to generate scientific conclusions, interpret findings, or produce original analytical content. All conceptual development, analysis, and interpretation remain the sole responsibility of the author, in accordance with journal policies on AI use disclosure.

3 Results

3.1 Overview of analytical findings

3.2 Typology of AI governance frameworks

3.2.1 Framework families

Table 1 summarizes the three principal families of AI governance frameworks identified across the reviewed literature, together with their core focus, strengths, limitations, and representative studies.

3.2.2.1 Cross-framework patterns

3.3 Lifecycle coverage analysis

3.3.1 Distribution across the AI lifecycle

Table 2 presents the distribution of governance coverage across the four stages of the AI lifecycle, highlighting the relative strengths and gaps identified in the reviewed literature.

3.3.2.1 Key observations

3.4 Operationalization gaps

3.4.1 Principle–implementation Gap

A persistent gap exists between articulated ethical principles and their institutional implementation. Frameworks such as UNESCO's ethical guidelines and similar principle-based models provide strong normative foundations but offer limited higher-education-specific operational guidance, particularly within procurement and audit workflows (Isaifan et al., 2025; Meimandi et al., 2025; Miah et al., 2025; NIST, 2023).

3.4.2 Sustainability integration gap

3.4.3 Contextual fragmentation

3.5 Synthesis of findings

3.6 Summary of key results

3.7 Operationalization of maqāṣid-based governance indicators

To address the identified principle–implementation gap, this study translates the maqāṣid al-sharīʿah framework into a set of measurable governance indicators aligned with key stages of the AI lifecycle. This operationalization aims to bridge the gap between ethical abstraction and institutional decision-making by specifying concrete metrics and evaluative criteria.

Table 3 presents a prototype mapping of maqāṣid objectives to governance indicators, associated metrics, and lifecycle stages. These indicators are designed to be adaptable across institutional contexts while maintaining conceptual alignment with value-based governance principles.

This operational mapping demonstrates how value-based principles can be embedded within institutional governance mechanisms through measurable indicators. Rather than treating maqāṣid as abstract ethical ideals, the framework integrates them into decision-making checkpoints across the AI lifecycle. For example, the preservation of intellect (ḥifẓ al-ʿaql) is operationalized through metrics related to cognitive dependency and critical thinking, while the preservation of property (ḥifẓ al-māl) is reflected in data governance and algorithmic fairness measures.

The proposed indicators are not intended as fixed standards but as adaptable benchmarks that institutions may refine according to their regulatory, cultural, and technological contexts. This approach supports both contextual flexibility and cross-institutional comparability, addressing a key limitation identified in existing governance frameworks.

Building on the governance indicators presented in Table 3, Table 4 presents the full Maqāṣid-Based Institutional AI Governance Scorecard, operationalized through a four-point maturity scale, weighted scoring criteria, and a Governance Maturity Index (GMI) with associated decision rules. To illustrate how the scorecard functions in practice, Table 5 provides a hypothetical application to an Islamic university operating within the Saudi Vision 2030 framework. This illustrative case is constructed for demonstration purposes and does not reflect data from any specific institution.

The scorecard operationalizes the five maqāṣid objectives into measurable governance indicators evaluated across four AI lifecycle stages (procurement, design, deployment, audit). Each indicator is scored on a four-point maturity scale; weighted scores are aggregated into an overall Governance Maturity Index (GMI) enabling institutional benchmarking and decision-making.

4 Discussion

The findings of this study address a persistent issue in the literature on artificial intelligence (AI) in higher education: the gap between ethical principles and their institutional operationalization. While prior research has advanced three major streams—SDG-aligned frameworks, ethical governance models, and value-based approaches—these streams remain insufficiently integrated into institution-level workflows and decision processes (Floridi et al., 2018; Jobin et al., 2019; Zawacki-Richter et al., 2019). Table 3 provides a concrete operational layer that directly addresses the principle–implementation gap identified in prior literature.

4.1 Interpreting the predominance of conceptual frameworks

As demonstrated in Table 3, this study provides an operational layer that directly addresses the principle–implementation gap identified across the reviewed literature. The results indicate that the substantial majority of existing frameworks remain predominantly conceptual, with limited empirical validation and weak mechanisms for institutional implementation. This finding is consistent with comparative syntheses showing that AI governance in higher education is still largely articulated at the level of normative principles rather than operational procedures (NIST, 2023; UNESCO, 2021; United Nations, 2015).

This pattern reflects the well-documented principle–implementation gap, whereby ethical commitments—such as fairness, transparency, and accountability—are defined at a high level but are not embedded within institutional decision-making processes, particularly in procurement, deployment, and audit workflows (OECD, 2019; UNESCO, 2021).

Rather than representing a failure, this predominance suggests that the field remains in a normative consolidation phase, where ethical consensus precedes operational maturity.

4.2 Fragmentation across governance levels and lifecycle stages

This fragmentation becomes more pronounced when examined across the AI lifecycle. As synthesized in this review, fewer than one-third of frameworks provide explicit guidance for procurement decisions, while post-deployment audit mechanisms are addressed in an estimated 10%–15% of reviewed frameworks (n ≈ 14–21), representing the most critically underrepresented stage across the entire lifecycle (OECD, 2019; Zawacki-Richter et al., 2019). Deployment-stage guidance, though more frequently present (est. ∼40%–50%; n ≈ 57–72), remains concentrated in general monitoring principles rather than structured implementation checkpoints, limiting end-to-end governance integration.

This uneven distribution indicates that governance is often concentrated in design and deployment stages, while critical decision points—such as system acquisition and long-term evaluation—remain weakly governed.

A critical insight emerging from the literature is that both Western and Islamic AI governance frameworks exhibit a similar structural limitation: the persistence of declarative ethical commitments without sufficient operational embedding. While Western frameworks emphasize principles such as fairness, transparency, and accountability, and Islamic models articulate values such as ʿadl, amānah, and iḥsān, both traditions often lack measurable implementation mechanisms. This convergence suggests that the challenge is not cultural but structural, thereby reinforcing the need for frameworks that translate ethical principles into actionable governance indicators.

4.3 Reassessing global governance frameworks

The findings also offer a more nuanced interpretation of global AI governance models. Frameworks such as UNESCO's ethical guidelines and NIST-informed approaches provide important normative and structural foundations. However, their application in higher education often remains limited at the level of institution-specific processes, particularly in procurement workflows, accountability structures, and lifecycle oversight (NIST, 2023; Selwyn, 2019).

This limitation reflects their cross-sectoral design rather than an inherent deficiency. In practice, their effectiveness depends on their translation into institutional procedures. The challenge, therefore, lies not in the absence of ethical frameworks, but in the absence of mechanisms that embed these frameworks within operational decision architectures.

Table 6 provides a systematic comparison of the proposed Maqāṣid-based framework with four established international AI governance models: the UNESCO Recommendation on the Ethics of Artificial Intelligence (UNESCO, 2021), the NIST AI Risk Management Framework (NIST, 2023), the OECD AI Principles (OECD, 2019), and the EU AI Act (European Union, 2024). The comparison identifies each framework's primary focus, institutional strengths, limitations within Islamic higher education contexts, and the specific contribution of the proposed framework. This mapping demonstrates how the Maqāṣid-based approach extends rather than replaces existing models, by introducing a values-translation layer that operationalizes ethical principles within lifecycle governance structures.

As Table 6 illustrates, the proposed framework occupies a complementary rather than competing position within the existing governance landscape. While UNESCO, NIST, OECD, and EU frameworks provide essential normative and risk-based foundations, they share a common structural limitation in the context of Islamic higher education: the absence of mechanisms for translating universal ethical principles into culturally grounded, institution-level governance processes. The Maqāṣid-based framework addresses this gap by introducing a values-translation layer — operationalized through the GMI scorecard (Tables 4, 5) — that maps Islamic ethical objectives to measurable governance indicators across the full AI lifecycle. This interoperability with established international models enhances rather than limits the framework's transferability, enabling institutions to align local governance practices with globally recognized standards while preserving cultural and epistemic coherence.

4.4 Reinterpreting value-based (maqāṣid) approaches

Value-based frameworks—particularly those grounded in maqāṣid al-sharīʿah—offer strong normative depth but remain underdeveloped in terms of institutional implementation (Hassan et al., 2022; Kamali, 2008; Zawacki-Richter et al., 2019).

4.5 From principles to institutional mechanisms

Building on these insights, the proposed framework advances the literature by shifting from normative articulation to operational design. It conceptualizes AI governance as a lifecycle-based system in which ethical principles are embedded within decision-making nodes.

Unlike models that treat ethics as a post hoc normative layer, this approach integrates ethical criteria directly into governance processes, thereby enhancing institutional accountability (NIST, 2023; Selwyn, 2019).

It is important to note that the framework is proposed as a prototype operational pathway (scorecard concept) rather than a fully validated model, pending further empirical testing.

4.6 Alignment with Saudi vision 2030 and national AI governance strategies

The findings of this study have direct implications for the strategic objectives of Saudi Vision 2030, particularly in the domains of digital transformation, human capital development, and institutional governance. As Saudi universities increasingly adopt AI technologies, the need for governance models that are both globally interoperable and locally grounded becomes critical.

The proposed Maqāṣid-based framework aligns closely with the Human Capability Development Program (HCDP) (Kingdom of Saudi Arabia, 2023), which emphasizes the development of a globally competitive and ethically grounded workforce. In this context, the preservation of intellect (ḥifẓ al-ʿaql) provides a conceptual and operational basis for promoting critical thinking, reducing cognitive dependency on AI systems, and supporting lifelong learning—key priorities within Vision 2030.

Similarly, the preservation of lineage and social integrity (ḥifẓ al-nasl) aligns with the Vision's emphasis on cultural identity and social cohesion, particularly in the face of rapid digital transformation. By embedding data privacy, identity protection, and safeguards against misuse (e.g., deepfake technologies), the framework contributes to maintaining societal trust and institutional legitimacy.

From a governance perspective, the framework is also consistent with the ethical principles articulated by the Saudi Data and Artificial Intelligence Authority (SDAIA, 2024), particularly in areas such as fairness, accountability, and transparency. The integration of Maqāṣid-based indicators within lifecycle governance processes allows institutions to operationalize these principles in a manner that is both contextually relevant and aligned with international standards such as the NIST AI Risk Management Framework.

Importantly, the proposed scorecard functions as a meso-level governance tool that bridges macro-level policy frameworks (e.g., SDGs, national AI strategies) with micro-level institutional practices. This directly addresses the fragmentation identified in the literature and provides a scalable mechanism for embedding ethical governance within university systems.

In this sense, the framework contributes to social sustainability by ensuring that AI integration in higher education remains culturally coherent, ethically grounded, and institutionally accountable. This is particularly significant in preventing the uncritical adoption of external technological models that may not align with local value systems, thereby supporting a form of sustainable and context-sensitive digital transformation.

4.7 Implications for sustainable higher education

The findings have broader implications for sustainable development in higher education (Leal Filho et al., 2019; Stephens et al., 2008). While many frameworks reference sustainability, the linkage to specific SDGs remains limited or implicit in most cases (de Silva, 2025; Floridi et al., 2018; Soko, 2025; United Nations, 2015).

By integrating lifecycle-based governance with value-based ethics, the proposed approach provides a pathway for aligning AI deployment with sustainability goals, particularly SDG 4 (Quality Education), SDG 9 (Innovation and Infrastructure), and SDG 16 (Institutional Accountability).

For Islamic higher education institutions—defined in this study as institutions whose institutional mission, governance mandates, or curricula are explicitly informed by Islamic ethical and epistemic frameworks, regardless of geographic location—this model supports contextual transferability rather than universal generalization, allowing adaptation across institutions with comparable value commitments.

4.8 Limitations and future research directions

This study has several limitations. First, it is based on an integrative literature review, and its conclusions are constrained by the scope of the reviewed corpus (2015–2026). Second, the proposed framework remains conceptual and requires empirical validation across diverse institutional contexts.

5 Conclusions

This integrative review synthesized 143 peer-reviewed studies (2015–early 2026) to examine how existing AI governance frameworks in higher education incorporate value-based and lifecycle-oriented ethical criteria. The findings indicate that dominant approaches—across SDG-aligned (Floridi et al., 2018; Jobin et al., 2019; United Nations, 2015; Zawacki-Richter et al., 2019), ethical-governance, and value-based traditions—remain largely conceptual, exhibit limited empirical validation, and provide uneven operational guidance across the AI lifecycle, with the most persistent gaps observed in procurement and post-deployment audit stages.

While global frameworks such as UNESCO's Recommendation on the Ethics of Artificial Intelligence, the NIST AI Risk Management Framework, and the EU AI Act (2024) (European Union, 2024) provide comprehensive normative and risk-based guidance (NIST, 2023; OECD, 2019; UNESCO, 2021), they offer limited explicit mechanisms for embedding context-specific ethical systems within institutional governance processes, particularly in culturally diverse or faith-based educational settings. In this regard, the proposed Maqāṣid al-Sharīʿah-informed framework aims to extend these approaches by introducing a values-translation layer that integrates ethical criteria into lifecycle-based governance structures across procurement, design, deployment, and audit (Dusuki and Abdullah, 2007).

The framework links the five higher objectives of Islamic ethics—preservation of religion, life, intellect, lineage, and property—to concrete governance functions. It proposes initial categories of measurable indicators, lifecycle checkpoints, and a prototype risk/decision matrix (scorecard) intended to support institutional assessment processes. However, these components are presented as conceptual artifacts pending empirical validation, including pilot implementation, inter-rater reliability testing, and expert-based (Delphi) evaluation.

It is important to distinguish between two types of claims advanced in this study. First, findings derived directly from the integrative literature review (2015–2026): dominant AI governance frameworks in higher education remain predominantly conceptual, exhibit uneven lifecycle coverage with persistent gaps in procurement and audit, and lack institution-specific operational mechanisms. Second, propositions advanced on the basis of these findings but requiring empirical validation: the maqāṣid-based governance indicators, the GMI scorecard, and the comparative framework mapping presented in Tables 3–6 represent prototype artifacts whose practical utility and reliability remain to be tested through pilot studies, Delphi consultation, and expert validation.

This study offers three potential contributions, pending empirical validation. First, it provides a theoretical contribution by operationalizing Maqāṣid al-Sharīʿah within contemporary AI governance discourse. Second, it contributes methodologically through the structured synthesis of interdisciplinary literature, identifying convergent governance patterns. Third, it advances a practical contribution by proposing a lifecycle-based governance model intended to inform institutional policy and practice.

From a policy perspective, the framework aligns with Saudi Vision 2030, particularly the Human Capability Development Program and the National Strategy for Data and AI (Auda, 2008; Kamali, 2008; Selwyn, 2019). The principle of ḥifẓ al-ʿaql supports critical thinking and lifelong learning, while ḥifẓ al-māl relates to equitable and responsible access to AI-enabled educational resources. The framework also aligns conceptually with SDAIA AI Ethics Principles and may facilitate interoperability between national regulatory standards and global governance models—for example, by mapping Maqāṣid principles to UNESCO's core values (human dignity, autonomy, justice) and to NIST's trustworthiness characteristics (validity, reliability, safety, fairness).

Despite its contributions, the study has several limitations. Methodologically, the review is restricted to peer-reviewed literature in English and Arabic, potentially excluding relevant scholarship in other languages and gray literature sources. Conceptually, the framework's grounding in Maqāṣid al-Sharīʿah may limit direct applicability in secular or multi-faith institutional contexts without adaptation. Temporally, the rapid evolution of generative AI technologies (2023–2026) introduces a dynamic that may require continuous framework refinement beyond the scope of this review. Most critically, the proposed framework and scorecard have not yet undergone empirical validation, and claims regarding operational effectiveness remain provisional.

Future research should prioritize empirical validation through mixed-methods pilot studies in specific institutional contexts, such as public universities in Saudi Arabia, private higher education institutions in the Gulf region, and Islamic universities in Southeast Asia. Key research questions include: (1) how institutional stakeholders perceive the feasibility and legitimacy of Maqāṣid-informed AI governance; (2) the extent to which the proposed scorecard demonstrates reliability and usability across evaluators; and (3) how the framework compares to established models (e.g., UNESCO Recommendation, NIST AI RMF, and EU AI Act risk classification) in terms of institutional adoption rates, governance maturity indicators, and stakeholder acceptance. Longitudinal studies examining governance evolution in the context of generative AI systems would provide further insights.

In conclusion, the proposed framework represents a shift from reactive, compliance-oriented AI governance toward value-centered and lifecycle-integrated institutional stewardship. While remaining conceptual pending empirical validation, it offers a structured and operationally grounded foundation for aligning technological innovation with sustainable, ethical, and context-sensitive educational objectives in an increasingly AI-driven higher education landscape.

References

• 1

  AflahahS.NisaK.AldeiaA. M. S.MarannuB.RamliM. (2026). AI, ethics, and islamic higher education: navigating the tensions between practical risk mitigation and religious normativity. J. Ilmiah Islam Futura26 (1), 1–24. 10.22373/jiifv26.i1.32978

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AlfalehM. (2026). Sustainable AI-driven assessment in higher education: a systematic review of fairness, transparency, pedagogical innovation, and governance. Sustainability18, 785. 10.3390/su18020785

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AudaJ. (2008). Maqāṣid al-Sharīʿah as Philosophy of Islamic Law: A Systems Approach. London: IIIT.

  • Google Scholar
  • View reference in article

• 4

  BarocasS.HardtM.NarayananA. (2019). Fairness and Machine Learning. Cambridge, MA: MIT Press.

  • Google Scholar
  • View reference in article

• 5

  BritchenkoI. (2025). Ethical AI in education: principles, governance, and responsible implementation. Pedagog. Educ. Manag. Rev.4 (22), 17–30. 10.36690/2733-2039-2025-4-17-30

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  ChenL.ChenP.LinZ. (2020). Artificial intelligence in education: a review. IEEE Access8, 75264–75278. 10.1109/ACCESS.2020.2988510

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  ChouT.-N. (2023). “Apply an integrated responsible AI framework to sustain the assessment of learning effectiveness,” in Proceedings of the International Conference on Computer Supported Education.

  • Google Scholar
  • View reference in article

• 8

  DakshitS.MokhtariK.KhalidA. (2026). Designing understandable and fair AI for learning: the PEARL framework for human-centered educational AI. Educ. Sci.16, 198. 10.3390/educsci16020198

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  de SilvaG. H. B. A. (2025). Data-Driven framework for aligning artificial intelligence with inclusive development in the Global South. Sustainability17 (21), 9360. 10.3390/su17219360

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  DusukiA. W.AbdullahN. I. (2007). Maqāṣid al-sharīʿah, maslahah, and corporate social responsibility. Am. J. Islamic Soc. Sci.24, 25–45.

  • Google Scholar
  • View reference in article

• 11

  European Union (2024). Regulation on artificial intelligence (AI act). Off. J. Eur. Union Regulation (EU) 2024/1689.

  • Google Scholar
  • View reference in article

• 12

  FloridiL.CowlsJ.BeltramettiM.ChatilaR.ChazerandP.DignumV.et al (2018). AI4People—an Ethical framework for a good AI society: opportunities, risks, principles, and recommendations. Minds Machines28, 689–707. 10.1007/s11023-018-9482-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  HashimR.LanggulungH. (2008). Islamic philosophy of education revisited. Intell. Disc.16, 1–16.

  • Google Scholar
  • View reference in article

• 14

  HassanM. K.KabirM. N.MuneezaA.MuhamadR. (2022). Islamic finance and artificial intelligence: a systematic literature review. J. Islamic Account. Bus. Res.13, 1–25.

  • Google Scholar
  • View reference in article

• 15

  HolmesW.BialikM.FadelC. (2019). Artificial Intelligence in Education: Promises and Implications for Teaching and Learning. Boston, MA: Center for Curriculum Redesign.

  • Google Scholar
  • View reference in article

• 16

  IEEE (2023). IEEE Standard for Algorithmic Bias Considerations (IEEE 2857). Piscataway, NJ: IEEE Standards Association.

  • Google Scholar
  • View reference in article

• 17

  IsaifanR.TawalbehA.HasnaM. (2025). From principles to practice: a novel matrix for evaluating AI-powered learning platforms based on the UNESCO ethical impact assessment tool. Front. Educ. 10:1640780. 10.3389/feduc.2025.1640780

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  JobinA.IencaM.VayenaE. (2019). The global landscape of AI ethics guidelines. Nat. Machine Intell.1, 389–399. 10.1038/s42256-019-0088-2

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  KamaliM. H. (2008). Maqāṣid al-Sharīʿah Made Simple. Herndon, VA: International Institute of Islamic Thought (IIIT).

  • Google Scholar
  • View reference in article

• 20

  KasarciI.Akın DemircanZ.Çeliker ErcanG.İnciT. (2025). Managing artificial intelligence ethics in higher education: a systematic framework for issues and policy recommendations. Int. J. Curr. Educ. Stud.4 (2), 112–137. 10.46328/ijces.223

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  Kingdom of Saudi Arabia (2016). Saudi Vision 2030. Riyadh: Government of Saudi Arabia. Available online at:https://www.vision2030.gov.sa (Accessed May 3, 2026)

  • Google Scholar
  • View reference in article

• 22

  Kingdom of Saudi Arabia (2023). Human Capability Development Program. Vision 2030 Realization Program. Riyadh: Government of Saudi Arabia.

  • Google Scholar
  • View reference in article

• 23

  Leal FilhoW.AzeiteiroU.AlvesF.PaceP.MifsudM.BrandliL.et al (2019). The role of higher education institutions in sustainability initiatives at the local level. J. Cleaner Prod.233, 1004–1015. 10.1016/j.jclepro.2019.06.059

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  LeonM. (2026). Responsible machine learning deployment: imperative framework for ethical action. WSEAS Transact. Comput. Res.14, 10.37394/232018.2026.14.12

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  LuckinR.HolmesW.GriffithsM.ForcierL. B. (2016). Intelligence Unleashed: an Argument for AI in Education. London: Pearson.

  • Google Scholar
  • View reference in article

• 26

  MadaniM.AgustianH.FaturahmanA.SutarmanA.IkhsanR. S. (2026). Sharīʿah-Guided artificial intelligence for ethical transformation in modern education. J. MENTARI, 10.33050/mentari.v4i2.975

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  MangalM.PardosZ. (2024). Implementing equitable and intersectionality-aware ML in education: a practical guide. Br. J. Educ. Technol., 10.1111/bjet.13484

  • CrossRef
  • Google Scholar

• 28

  ManganelloF.NicoA.RagusaM.BoccuzziG. (2025). Testing the applicability of a governance checklist for high-risk AI-based learning outcome assessment in Italian universities under the EU AI act annex III. Front. Artif. Intell., 10.3389/frai.2025.1718613

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  MeimandiK.ReuelA.Aranguiz-DiasG.RahamaH.AyadiA. (2025). “An adaptive responsible AI governance framework for decentralized organizations,” in Proceedings of the AAAI/ACM Conference on AI (Ethics, and Society. 10.1609/aies.v8i2.36633

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  MiahM. N. I.UddinM. J.AhmedM. W. (2025). Regulating artificial intelligence in education: analyzing legal and ethical frameworks for the deployment of AI and machine learning models in U. S. educational institutions. J. Comput. Sci. Technol. Stud.7 (11), 387–404. 10.32996/jcsts.2025.7.11.37

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  National Institute of Standards and Technology (NIST) (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). Gaithersburg. MD: NIST.

  • Google Scholar
  • View reference in article

• 32

  NilamsariF.IchsanW.AsasunnajaR.SabrinaF. (2025). Etika penggunaan ChatGPT dalam pendidikan dan ekonomi islam: tinjauan maqāṣid al-sharīʿah. Cyberspace9 (2), 50–63. 10.22373/cj.v9i2.31834

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  OECD (2019). OECD Principles on Artificial Intelligence. Paris: OECD Publishing.

  • Google Scholar
  • View reference in article

• 34

  Partnership on AI (2023). Algorithmic Impact Assessments: A Practical Framework for Public Agency Accountability. New York, NY: Partnership on AI. Available online at:https://partnershiponai.org (Accessed April 1, 2026).

  • Google Scholar
  • View reference in article

• 35

  RussellS. (2019). Human Compatible. New York, NY: Viking.

  • Google Scholar
  • View reference in article

• 36

  Saudi Data and Artificial Intelligence Authority (SDAIA) (2024). AI Ethics Principles for the Kingdom of Saudi Arabia. Riyadh: SDAIA.

  • Google Scholar
  • View reference in article

• 37

  SelwynN. (2019). Should Robots Replace Teachers? AI and the Future of Education. Cambridge, UK: Polity Press.

  • Google Scholar
  • View reference in article

• 38

  SokoJ. (2025). Ethical concerns and institutional policy responses to artificial intelligence (AI) in higher education across sub-Saharan Africa. J. Kenya Nat. Commission UNESCO. 5 (1). 10.62049/jkncu.v5i1.426

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  StephensJ. C.HernandezM. E.RománM.GrahamA. C.ScholzR. W. (2008). Higher education as a change agent for sustainability in different cultures and contexts. Int. J. Sustain. High. Educ.9, 317–338. 10.1108/14676370810885916

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  SwistT.Buckingham ShumS.GulsonK. N. (2024). Co-producing AIED ethics under lockdown: an empirical study of deliberative democracy in action. Int. J. Artif. Intell. Educ.34, 670–705. 10.1007/s40593-023-00348-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  UNESCO (2021). Recommendation on the Ethics of Artificial Intelligence. Paris: UNESCO.

  • Google Scholar
  • View reference in article

• 42

  United Nations (2015). Transforming Our World: The 2030 Agenda for Sustainable Development. New York, NY: United Nations.

  • Google Scholar
  • View reference in article

• 43

  WeiZ.LiuZ.WangT.YongzhenL. (2025). Ethical risks and point-of-use governance of AI in IPE: insights from interviews with 17 instructors. Critical Humanistic Social Theory2 (3). 10.62177/chst.v2i3.599

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  WinfieldA. F.JirotkaM. (2018). Ethical governance is essential to building trust in robotics and artificial intelligence systems. Philos. Transact. R. Soc. A376, 20180085. 10.1098/rsta.2018.0085

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  Zawacki-RichterO.MarínV. I.BondM.GouverneurF. (2019). Systematic review of research on artificial intelligence applications in higher education. Int. J. Educ. Technol. High. Educ.16, 39. 10.1186/s41239-019-0171-0

  • CrossRef
  • Google Scholar
  • View reference in article