Algorithmic dependence and digital colonialism: a conceptual framework for artificial intelligence in education and knowledge systems of the Global South

Ahmed, Samar A.

doi:10.3389/feduc.2026.1720563

OPINION article

Front. Educ., 21 January 2026

Sec. Digital Learning Innovations

Volume 11 - 2026 | https://doi.org/10.3389/feduc.2026.1720563

Algorithmic dependence and digital colonialism: a conceptual framework for artificial intelligence in education and knowledge systems of the Global South

Samar A. Ahmed^1,2^*

¹Faculty of Medicine, Ain Shams University, Cairo, Egypt
²Rabdan Academy, Abu Dhabi, United Arab Emirates

1 Introduction

AI's integration into education has accelerated across all regions, driven by ambitions of efficiency, personalization, and expanded access. International organizations such as UNESCO (2021) and the OECD (2023) increasingly frame AI as a mechanism for educational equity and knowledge democratization. However, these narratives often obscure a parallel development: the consolidation of AI infrastructures—including training data, model architectures, analytics platforms, and academic publishing systems—within a small number of Global North corporations and states (Crawford, 2021). This concentration produces dependencies that extend beyond technical adoption into the epistemic foundations of education.

Across the Global South, universities and education systems increasingly rely on proprietary AI-enabled platforms such as Turnitin for plagiarism detection, Elsevier's Scopus for research analytics, and Google or Microsoft cloud ecosystems for teaching and learning. These systems are developed within specific epistemic, linguistic, and regulatory contexts, embedding assumptions about originality, relevance, quality, and authority. For example, Scopus indexing practices privilege English-language journals and Global North publication venues, systematically marginalizing regional scholarship (Sangwa et al., 2025). In this way, AI-mediated infrastructures shape not only how education is delivered, but also which forms of knowledge are rendered visible, legitimate, and valuable.

This paper adopts a conceptual-analytical approach rather than an empirical research design. Its aim is not to measure the effects of AI adoption, but to systematize existing critical scholarship into an integrated framework capable of explaining how algorithmic dependence emerges and persists within education systems of the Global South. Building on critical studies of digital colonialism and AI governance, the paper proposes a conceptual framework of AI-driven digital colonialism structured around four interrelated dimensions: data colonialism, infrastructure dependence, epistemic colonialism, and governance colonialism. It addresses two guiding questions: (1) How does AI reinforce digital colonialism in Global South education systems? (2) What conceptual dimensions are required to analyze algorithmic dependence in education?

2 Literature review

2.1 Digital colonialism and data colonialism

Digital colonialism has been defined as the domination of global digital ecosystems by a limited number of corporations and states that control software platforms, data infrastructures, and information flows (Kwet, 2019). Extending this analysis, Couldry and Mejias (2019) introduce the concept of data colonialism, framing large-scale data extraction as a continuation of historical colonial logics of appropriation. Rather than land or labor, data colonialism treats human activity itself as a raw material for accumulation.

Within education, data colonialism manifests through learning management systems, analytics dashboards, biometric proctoring technologies, and AI-driven assessment tools that convert pedagogical activity into extractable and monetizable data. Empirical studies demonstrate that education systems across Africa and Latin America rely disproportionately on U.S.- or China-based digital infrastructures, resulting in limited local control over educational data (Hassan, 2022). While this literature establishes data extraction as a structural condition of contemporary digital systems, it rarely theorizes its specific implications for educational sovereignty or institutional autonomy.

2.2 Artificial intelligence in education

AI applications—ranging from adaptive tutoring systems to predictive analytics and automated assessment—have become central to education reform agendas worldwide (Holmes et al., 2021; Williamson and Eynon, 2020). Much of the AI-in-education literature emphasizes innovation, scalability, and efficiency gains. However, critical scholarship highlights uneven distribution of benefits, contextual misalignment, and governance asymmetries, particularly in the Global South.

Studies from Latin America document heavy reliance on imported AI tools and English-language datasets, limiting meaningful localization and cultural relevance (Cobo, 2022). In Southeast Asia, national AI strategies prioritize economic competitiveness while paying limited attention to data sovereignty or epistemic implications (ASEAN, 2024). These findings suggest that AI adoption often proceeds without parallel investments in infrastructural autonomy, epistemic diversity, or locally grounded governance, thereby reinforcing dependency rather than fostering capacity-building.

2.3 Epistemic Injustice and Knowledge Systems

The concept of epistemic injustice (Fricker, 2007) and de Santos (2014) notion of epistemicide provide a theoretical foundation for analyzing how dominant knowledge systems marginalize alternative epistemologies. Recent scholarship extends these ideas to algorithmic systems, demonstrating how AI reproduces visibility hierarchies embedded in its training data and design assumptions (Noble, 2018; Crawford, 2021).

Empirical research in India and Brazil shows that AI-based translation, grading, and recommendation systems systematically underperform when applied to regional languages, culturally situated discourse, and non-Western rhetorical traditions (Sharma and Balaji, 2023; Oliveira and Cruz, 2024). Although this body of work clearly documents epistemic harm, it remains fragmented across disciplines and is seldom integrated into analyses of AI in formal education systems.

2.4 Explicitly identified gap

Taken together, the literature reveals three persistent limitations. First, critical studies of digital and data colonialism rarely focus explicitly on education as a distinct institutional domain. Second, AI-in-education research often neglects colonial power relations and questions of sovereignty. Third, analyses of epistemic injustice are insufficiently connected to technological infrastructure and governance arrangements.

This paper addresses these gaps by synthesizing insights from digital colonialism, AI-in-education, and epistemic justice literatures into a single conceptual framework that explains how AI-driven systems simultaneously reproduce data extraction, infrastructural dependence, epistemic marginalization, and governance asymmetry within Global South education systems.

3 Conceptual framework: AI-driven digital colonialism in education

The conceptual framework is inductively derived from recurring patterns identified across the reviewed literature. Each dimension reflects a distinct mechanism repeatedly documented in empirical and theoretical studies, rather than an abstract or speculative categorization.

3.1 Data colonialism

Educational data—including learning analytics, behavioral logs, biometric proctoring data, and research performance metrics—are routinely extracted by global vendors headquartered in the Global North. For example, South African universities using U.S.-based online proctoring software have raised concerns regarding the storage and processing of biometric data outside national jurisdictions (Mutimukwe et al., 2025). Such practices exemplify what Couldry and Mejias (2019) describe as “data relations of production,” in which local users function primarily as sources of value for external actors.

3.2 Infrastructure dependence

Dependence on proprietary AI ecosystems creates technological lock-in and constrains institutional autonomy. Universities across Latin America increasingly rely on Microsoft Azure and Google Cloud to host AI-enhanced learning environments, often in the absence of viable local alternatives (Cobo, 2022). This dependence inflates long-term costs, limits data localization, and reduces bargaining power. In response, the African Union's Digital Transformation Strategy (2024) explicitly identifies AI sovereignty as a strategic objective to reduce reliance on foreign infrastructure.

3.3 Epistemic colonialism

AI systems trained primarily on Western corpora encode cultural and linguistic biases into curriculum design, assessment practices, and knowledge recommendation. Automated essay scoring systems undervalue non-Western rhetorical traditions, while algorithmic recommendation engines disproportionately surface Western journals and authors. Studies examining Indonesia's AI curriculum alignment reveal a persistent mismatch between imported AI content and indigenous pedagogical philosophies (Raharjo and Rohmadi, 2025). These processes sustain epistemic hierarchies in which Western knowledge systems define global academic norms (de Santos, 2014).

3.4 Governance colonialism

Governance colonialism occurs when ethical, legal, and policy frameworks governing AI are imported wholesale without contextual adaptation. Many Global South countries adopt OECD or EU “trustworthy AI” principles despite distinct cultural, educational, and regulatory environments (Organisation for Economic Co-operation and Development (OECD), 2023). Although UNESCO's Beijing Consensus (UNESCO, 2019) advocates cultural sensitivity, implementation often defaults to Northern templates. Emerging initiatives such as Africa's AI for Development Network and Latin America's AI Observatory (OECD.AI, 2024) indicate early efforts to articulate regionally grounded governance perspectives.

3.5 Interconnected dynamics

These four dimensions are mutually reinforcing rather than discrete. Data extraction feeds infrastructure dependence; infrastructure dependence amplifies epistemic effects; and governance importation stabilizes the overall system of algorithmic control. The framework thus responds directly to calls in the literature for integrative models capable of explaining AI's systemic educational impacts rather than isolated technological applications.

4 Discussion

4.1 Consequences of AI-driven digital colonialism

Drawing on documented cases and empirical findings from the reviewed literature, the framework reveals four recurring consequences. Epistemic dependency emerges as AI tools privilege Northern knowledge bases and discourse norms (Selwyn, 2019). Innovation lock-in results from deep integration with hyperscale cloud platforms, restricting experimentation and local development (Cobo, 2022; Kwet, 2019). Educational inequality widens as Global North institutions develop customized AI systems while Global South institutions rely on generic tools with limited localization (Sharma and Balaji, 2023). Finally, policy displacement occurs when imported governance frameworks misalign with local pedagogical values, as observed in parts of Southeast Asia and Africa (ASEAN, 2024).

4.2 Counter-strategies for educational sovereignty

Open-source AI initiatives such as BLOOM (BigScience, 2022) and AfricaNLP (Hassan, 2022) demonstrate the potential for multilingual and culturally inclusive AI development. Regional collaborations among the African Union, ASEAN, and MERCOSUR foster South–South cooperation and shared infrastructure development (OECD.AI, 2024). Indigenous data governance frameworks, including Kenya's Data Protection Act and Brazil's LGPD, support local control over educational data. Together, these strategies suggest that sovereignty-oriented approaches require coordinated action across data, infrastructure, epistemology, and governance.

4.3 Toward a pluralistic AI future

AI need not reproduce colonial dependencies. When governed through open, context-sensitive, and epistemically plural frameworks, AI can support rather than undermine educational autonomy. The framework advanced here offers a basis for evaluating whether AI deployments move education systems toward sovereignty or deeper dependency.

5 Conclusion

AI's integration into education promises innovation but risks reinforcing structural inequities. This paper has developed and justified a conceptual framework—data colonialism, infrastructure dependence, epistemic colonialism, and governance colonialism—derived from critical scholarship to analyze algorithmic dependence in Global South education systems. By making explicit the mechanisms through which AI reproduces colonial relations, the framework provides an analytical tool for future empirical research, comparative analysis, and policy design. Counter-strategies are emerging, yet their effectiveness depends on whether education systems move beyond adoption toward sovereignty-oriented governance.

Author contributions

SA: Writing – original draft, Conceptualization, Formal analysis, Methodology, Investigation, Writing – review & editing.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The author SA declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. Generative AI was used for the revision of text for academic writing and review of citations.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

African Union's Digital Transformation Strategy. (2024). African Union Digital Transformation Strategy for Africa (2024–2030). Addis Ababa: African Union Commission.