Saudi Arabia Built a National AI Governance Machine. Its Universities Are Still Catching Up.

Saudi Arabia has spent nearly a decade constructing one of the world’s most ambitious national frameworks for artificial intelligence governance, anchored by the Saudi Data and Artificial Intelligence Authority, or SDAIA, and driven by the sweeping digital transformation agenda of Vision 2030. From the AI Ethics Principles issued in 2023 to the Generative AI Guidelines published in 2024 and 2025, the Kingdom has produced a layered architecture of ethical principles, risk classifications, operational guidance, and occupational standards designed to govern how artificial intelligence is developed, deployed, and used across government, industry, and education. Yet a new study published in Frontiers in Education reveals a striking discontinuity at the very point where this national machinery meets the classroom: while the national governance landscape is detailed and increasingly operational, the translation of those principles into university-specific policies remains uneven, under-specified, and, in many cases, publicly invisible.

The research, conducted by Fouad Ahmed Atallah of Jouf University’s College of Sharia and Law, employed systematic documentary analysis combined with gap analysis to examine eight national-level policy and strategic documents issued by SDAIA and the Government of Saudi Arabia. The corpus included the Principles and Controls of AI Ethics, the Generative AI Guidelines in both government and public versions, the AI Adoption Framework, the Saudi Academic Framework for AI Qualifications, the National Occupational Standard Framework for Data and AI, the National Strategy for Data and AI, and the Vision 2030 education pillar. Each document was analyzed through a four-step procedure involving descriptive content analysis, neo-institutional coding, Maqasid al-Shariah assessment, and gap analysis, allowing the researcher to identify not only what the documents say but also what they fail to specify for the distinctive organizational structures of higher education.

The findings reveal a national governance architecture that is both substantial and increasingly operational. The Principles and Controls of AI Ethics articulates seven core principles—Fairness, Privacy and Security, Humanity, Social and Environmental Benefits, Reliability and Safety, Transparency and Explainability, and Accountability and Responsibility—within a risk-based governance approach that assigns organizational responsibilities across the AI system lifecycle. The Generative AI Guidelines complement these principles with practical guidance on data handling, human oversight, accountability for AI-assisted outputs, and alignment with the Personal Data Protection Law and National Cybersecurity Authority standards. The AI Adoption Framework specifies five interconnected dimensions of organizational readiness, while the National Occupational Standard Framework defines sixteen occupational standards for AI-related roles, including AI Ethicist and Chief AI Officer, signaling national expectations that organizations will maintain dedicated governance personnel rather than treating ethics as an ancillary function.

Yet when Atallah examined how these national provisions are reflected in publicly accessible university-level governance documents, the picture changed dramatically. A systematic audit of seventeen purposively selected Saudi universities, conducted on 1 June 2026 through official institutional websites and policy repositories using predefined Arabic and English search terms, identified publicly accessible formal AI governance policies or guidelines at only six institutions: King Khalid University, King Saud University, King Abdulaziz University, Princess Nourah Bint Abdulrahman University, Jazan University, and Taif University. For the remaining eleven universities, no comparable publicly accessible document was identified through the documented search procedure. The identified institutional documents varied considerably in scope and form, ranging from university-wide AI policies to generative AI guidance specifically for faculty and students, and from governance frameworks addressing organizational accountability to policies focused on e-learning contexts.

The study is careful to distinguish between the absence of publicly accessible documentation and the absence of internal governance arrangements. A negative search outcome refers only to publicly observable policy availability and should not be interpreted as evidence that a university lacks internal, unpublished, restricted-access, informal, or developing AI governance mechanisms. Nevertheless, the audit demonstrates substantial heterogeneity in the publicly documented institutional translation of national AI governance frameworks across the selected cases, a pattern that neo-institutional theory helps to contextualize. Universities operate under interacting coercive, mimetic, and normative pressures that need not point in the same direction, and when organizations face competing institutional expectations without established mechanisms for reconciling them, neo-institutional theory identifies decoupling as one possible organizational response in which formal structures become only loosely connected to actual practice.

A second critical finding concerns the relationship between Saudi national AI governance and Islamic ethical frameworks. The national documents reference human dignity, fairness, responsibility, cultural values, and alignment with the Kingdom’s regulatory environment—provisions that may be analytically compatible with Islamic ethical principles. However, the study found that the documents do not systematically operationalize Maqasid al-Shariah, the higher objectives of Islamic law, as a formal framework for AI governance. The five essential protections traditionally associated with Maqasid al-Shariah—protection of religion, life and dignity, intellect, lineage and family, and property—offer potential normative lenses for examining AI-related concerns including misinformation, privacy, algorithmic bias, human autonomy, and economic harm. Related concepts such as amanah, meaning trusteeship and responsibility, and adl, meaning justice, may likewise provide normative resources for accountability and fairness in AI governance. Yet the documentary evidence establishes analytical correspondence rather than explicit operationalization, distinguishing what the study interprets from what the documents themselves prescribe.

This distinction matters because it identifies an opportunity for future policy development rather than an established deficiency. The gap analysis revealed that national frameworks provide comparatively limited higher-education-specific guidance on how broad AI governance principles should be translated into university governance structures, academic integrity systems, faculty and student responsibilities, assessment design, and institutional monitoring mechanisms. Universities possess organizational characteristics that distinguish them from many other institutions, including faculty governance, disciplinary autonomy, research ethics structures, student assessment processes, and decentralized academic decision-making. National frameworks leave universities to determine how responsibilities, oversight mechanisms, and accountability should operate within their own institutional contexts, creating a policy-to-institution translation challenge rather than an absence of national governance.

The study’s implications extend beyond the Saudi context to other Arab and Muslim-majority higher education systems seeking to align emerging global AI governance standards with local Islamic and cultural values. More explicit engagement with Maqasid al-Shariah, amanah, and adl could provide culturally grounded normative resources for institutional AI governance without requiring a single standardized interpretation, provided that their application remains attentive to scholarly plurality and institutional context. The Saudi case suggests that bridging national governance architecture and university-level policy development may require a combination of higher-education-specific translation guidance, institutional capacity building, faculty participation, adaptable governance models, peer learning, and context-sensitive ethical frameworks.

For teaching and learning specifically, the findings point to the importance of institutional guidance that provides sufficient clarity for faculty and students while allowing appropriate adaptation across disciplines. Assessment design is particularly important because generative AI challenges conventional assumptions about authorship and independent student work, and institutional guidance could encourage approaches that emphasize demonstrated reasoning, critical evaluation of AI-generated content, documentation of learning processes, and appropriate disclosure of AI assistance. Faculty development should extend beyond technical familiarity with generative AI tools to address AI literacy, ethical and responsible use, assessment redesign, data privacy, intellectual property, bias, and verification of AI-generated information. Students likewise require explicit guidance to understand the capabilities and limitations of generative AI, distinguish legitimate academic assistance from misconduct, and critically evaluate AI-generated outputs.

The study acknowledges several methodological limitations. The institutional audit was restricted to publicly accessible documents, the national documentary corpus reflects materials identified and reviewed between March and May 2026, and the qualitative documentary analysis was conducted by a single analyst. Future research employing multiple independent coders, broader stakeholder samples, internal institutional records, interviews, and comparative case studies would strengthen analytical reliability and help determine how national AI governance principles are translated into university practice, which institutional factors shape that process, and how culturally grounded ethical frameworks can contribute to responsible AI governance in higher education. The Saudi case ultimately demonstrates that building a national AI governance machine is one thing; ensuring that its principles reach the lecture hall, the assessment rubric, and the research lab is quite another.

Subject of Research: Institutional policies for responsible generative AI use in Saudi higher education: a policy review and gap analysis of national-to-institutional implementation

Article Title: Institutional policies for responsible generative AI use in Saudi higher education: a policy review and gap analysis of national-to-institutional implementation

Article References: Atallah, F. A. (2026). Institutional policies for responsible generative AI use in Saudi higher education: a policy review and gap analysis of national-to-institutional implementation. Frontiers in Education, 11, Article 1917686. https://doi.org/10.3389/feduc.2026.1917686

Image Credits: AI Generated

DOI: 10.3389/feduc.2026.1917686

Keywords: Institutional, policies, responsible, generative, Saudi, higher, education, policy, review, analysis, national-to-institutional, implementation