AI in Education: Personalised Learning for the MENA Region

When the COVID-19 pandemic forced 1.6 billion students worldwide out of classrooms in 2020, the Middle East experienced an education technology revolution compressed into weeks rather than years. Governments that had been cautiously piloting digital learning suddenly deployed national-scale platforms, while private schools scrambled to maintain instructional continuity through whatever means available. Three years later, the emergency has passed but its legacy endures: a region transformed in its openness to technology-enabled education and increasingly focused on artificial intelligence as the next frontier. From adaptive learning platforms that adjust to individual student needs to AI tutors that provide personalised support at scale, the MENA education sector is exploring applications that could address longstanding challenges of educational quality, access, and relevance.

The stakes for educational transformation in the region are substantial. Despite significant investment in education infrastructure over recent decades, learning outcomes across much of the MENA region remain below potential. World Bank analysis has documented what researchers term the “learning crisis”—a gap between school enrollment, which has expanded dramatically, and actual learning, which has improved far less. The average student in the MENA region scores below global benchmarks on international assessments despite often attending school for as many years as peers in higher-performing systems. This learning gap carries economic consequences: research on human capital suggests that quality-adjusted educational attainment in the region lags quantity-based measures substantially, limiting the workforce productivity that economic diversification strategies require. AI-enabled personalised learning offers potential to address this quality challenge by adapting instruction to individual student needs at scale—a capability that human teachers, however skilled, cannot provide when facing classrooms of thirty or more students.

The regional education technology market has grown correspondingly. According to HolonIQ analysis, the MENA edtech sector attracted over $500 million in investment between 2020 and 2023, with AI-enabled platforms capturing an increasing share of capital flows. Dubai-based Alef Education, which uses AI to personalise K-12 learning pathways, raised $515 million in a landmark funding round that valued the company at over $1 billion—the largest edtech investment in the region’s history. Saudi Arabia’s Noon Academy has built a platform combining live tutoring with AI-driven practice recommendations, growing to serve over 12 million students across Arab markets. Jordan’s Abwaab offers personalised exam preparation for secondary students, using AI to identify knowledge gaps and recommend targeted practice. These companies represent regional responses to global educational challenges, developing solutions that account for Arabic language requirements, regional curriculum standards, and cultural contexts that global platforms often overlook.

The Science and Application of Personalised Learning

The premise underlying AI-enabled personalised learning rests on well-established educational research: students learn more effectively when instruction adapts to their current knowledge, learning pace, and individual needs. The concept of the “zone of proximal development”—the space between what a learner can do independently and what they can achieve with appropriate support—has guided educational practice since Vygotsky articulated it nearly a century ago. Traditional classroom instruction necessarily compromises this principle, targeting the middle of diverse student populations while leaving some students under-challenged and others overwhelmed. AI systems can potentially resolve this compromise by providing each student with content and support calibrated to their individual zone, advancing as they demonstrate mastery and providing additional scaffolding where they struggle. The theoretical promise has attracted enormous investment, though the evidence base for AI-enabled personalisation remains mixed and contested.

The most rigorously evaluated personalised learning implementations show meaningful but modest effects. A large-scale study by the RAND Corporation examining personalised learning approaches in US schools found positive effects on mathematics achievement equivalent to roughly three additional weeks of learning—statistically significant but not transformational. The National Bureau of Economic Research has documented positive effects from adaptive learning in higher education, with students in adaptive courseware sections demonstrating stronger outcomes than those in traditional formats. However, these studies also reveal substantial implementation variation: the same platforms produce dramatically different results depending on how teachers integrate them, how students engage with them, and how well the content aligns with learning objectives. Technology alone does not guarantee improvement; its effects depend on the broader instructional context into which it is introduced. This finding has important implications for MENA implementations, where the temptation to deploy technology as a solution in itself—without corresponding investment in teacher training, curriculum alignment, and pedagogical integration—risks reproducing disappointing results.

Arabic language AI presents particular challenges that affect personalised learning applications in the region. Natural language processing for Arabic lags significantly behind English, reflecting both the relative scarcity of Arabic training data and the language’s morphological complexity—a single Arabic word can contain as much information as an English phrase, complicating the analysis that underlies adaptive learning systems. Research presented at the Association for Computational Linguistics has documented these performance gaps across multiple NLP tasks. For educational applications, these limitations affect everything from reading comprehension assessment to conversational tutoring to automated essay scoring. Regional edtech companies have invested heavily in Arabic AI capabilities—Alef Education has built proprietary Arabic NLP systems, while G42’s collaboration with academic institutions aims to develop foundation models for Arabic that could benefit educational applications. Progress is accelerating, but current Arabic AI capabilities constrain what personalised learning systems can accomplish compared to English-language counterparts, particularly for applications requiring nuanced language understanding.

Implementation Realities Across the Region

Government-led initiatives represent the largest-scale AI in education deployments across MENA, though their outcomes remain difficult to assess given limited independent evaluation. The UAE’s Ministry of Education has partnered with Alef Education to deploy AI-enabled learning across public schools, with reported implementation reaching over 500,000 students. Saudi Arabia’s Madrasati platform, launched during the pandemic, has incorporated AI elements for content recommendation and learning pathway personalisation, though the system serves primarily as a learning management system rather than a fully adaptive platform. Qatar Foundation’s initiatives have explored AI for personalised learning in mathematics, with pilot programmes reporting improved student engagement and outcomes. Egypt has announced ambitious plans for AI in education as part of its digital transformation strategy, though implementation progress has been slower than initial announcements suggested. These government programmes share common challenges: deploying technology at national scale, ensuring equitable access across diverse student populations, building teacher capacity to leverage new tools effectively, and demonstrating outcomes that justify continued investment.

The private sector has moved more quickly, particularly in the Gulf states where parents with high expectations and substantial resources drive demand for educational innovation. International schools across Dubai and Abu Dhabi have implemented adaptive learning platforms in mathematics, literacy, and test preparation, often supplementing rather than replacing traditional instruction. Private tutoring—a major industry across the MENA region—has been disrupted by AI-enabled platforms that offer personalised practice at lower cost than human tutors, though the highest-stakes examination preparation still commands premium pricing for human expertise. Higher education institutions have begun exploring AI for personalised learning support, with the American University of Sharjah piloting adaptive systems in foundational courses and the King Abdullah University of Science and Technology developing AI tutoring for technical disciplines. These private sector implementations benefit from greater flexibility than government programmes and often serve as proving grounds for approaches that might eventually scale to public education—though the risk that innovation remains concentrated among privileged populations raises equity concerns that policymakers must address.

Teacher perspectives on AI in education reveal both enthusiasm and anxiety that implementations must navigate carefully. Surveys across the region indicate that educators generally recognise AI’s potential to reduce administrative burden, provide better diagnostic information about student learning, and offer supplementary support that extends teaching capacity. However, concerns about job displacement, loss of professional autonomy, and the dehumanisation of education emerge consistently. A Teaching and Learning International Survey analysis found that teachers in MENA countries report lower confidence in their ability to integrate technology effectively compared to global averages—a gap that reflects both infrastructure limitations and professional development needs. Successful implementations have addressed teacher concerns through transparent communication about AI’s role as augmentation rather than replacement, substantive professional development that builds capacity and confidence, and involvement of teachers in implementation decisions that affect their practice. Programmes that impose AI systems on teachers without such engagement consistently struggle to achieve adoption and impact.

Challenges, Risks, and the Path Forward

Data privacy concerns present significant challenges for AI-enabled education, particularly given the sensitive nature of information about children’s learning and development. Effective personalisation requires detailed data about student performance, behaviour, and progress—data that, if misused, could affect students’ opportunities and life trajectories. The MENA region’s regulatory frameworks for educational data protection remain underdeveloped compared to jurisdictions like the European Union, where GDPR imposes strict requirements on processing children’s data. Parents across the region express concern about data collection and use, even when broadly supportive of educational technology. The UNICEF guidance on AI for children provides principles that educational deployments should satisfy, including data minimisation, purpose limitation, and meaningful transparency to families. Organisations implementing AI in education must develop data governance frameworks that protect student privacy while enabling the personalisation that drives educational value—a balance that requires careful attention to technical safeguards, policy frameworks, and stakeholder communication.

Equity concerns demand particular attention given existing educational inequalities across and within MENA countries. The risk that AI-enabled personalised learning benefits primarily students who already enjoy educational advantages—those with reliable internet access, supportive home environments, and schools with capacity to implement new technologies effectively—would widen rather than narrow achievement gaps. Evidence from other technology implementations supports this concern: digital learning often produces larger benefits for advantaged students who have the literacy, motivation, and support to leverage new tools effectively. Addressing equity requires deliberate strategies: ensuring infrastructure investments reach underserved communities, designing platforms that function effectively with limited bandwidth and older devices, developing content that reflects diverse cultural contexts, and providing implementation support that helps under-resourced schools achieve effective adoption. National strategies that celebrate technological ambition without addressing equity dimensions risk creating education systems that are innovative in their best schools while leaving struggling schools further behind.

The path forward for AI in MENA education requires balancing ambition with realism about what technology can accomplish and what implementation demands. The most effective approaches will likely combine AI capabilities with human expertise rather than attempting to replace teachers with technology—recognising that education involves relational, motivational, and developmental dimensions that AI cannot replicate. Success will require sustained investment not only in technology but in teacher development, curriculum alignment, and organisational change management that enable effective integration. Measurement and evaluation must improve, with governments and organisations investing in rigorous assessment of what works rather than relying on vendor claims or anecdotal success stories. And regional collaboration can accelerate progress, with countries learning from each other’s experiences and sharing resources for Arabic language AI development that benefits all. The opportunity to transform education through AI is genuine; realising that opportunity will require strategy, investment, and persistence commensurate with the challenge.