Jake Okechukwu Effoduh and Miracle Okumu Mudeyi interrogate the coloniality of AI, the extractive political economy of data, and the structural inequalities embedded in algorithmic decision-making, arguing for a decolonial approach to AI policy to escape a cycle of technological dependency.
TWAILR: Reflections ~ 74/2025
Introduction
In a dimly lit internet café in Nairobi, a group of university students huddle around a borrowed laptop, training an artificial intelligence (AI) model with limited computational power. Halfway across the world, a Silicon Valley startup backed by millions in venture capital fine-tunes an advanced algorithm on one of the world’s most powerful supercomputers. The contrast is stark: one side grapples with infrastructural and financial constraints, while the other benefits from an entrenched ecosystem of technological dominance. Yet both claim to be shaping the future of AI. But whose future is being written, and whose interests does it serve?
Artificial intelligence is rapidly becoming the backbone of global governance, economic development, and security. Yet, as with previous technological revolutions, the Global South finds itself largely positioned as a consumer rather than a producer of AI-driven innovations. Third World Approaches to International Law (TWAIL) scholars have long argued that international law is not a neutral force but an instrument that sustains global hierarchies rather than dismantling them. This dynamic is particularly evident in AI governance, where regulatory frameworks are shaped by the priorities of the Global North, leaving the Third World with little agency in defining the rules of engagement.
This reflection critically examines AI governance through a TWAIL lens, interrogating the coloniality of AI, the extractive political economy of data, and the structural inequalities embedded in algorithmic decision-making. It argues that without a decolonial approach to AI policy, the Global South risks being further entrenched in a cycle of technological dependency, perpetuating the very asymmetries that TWAIL scholarship seeks to expose and challenge.
The Persistence of the Empire and Imperialism: The Coloniality of AI
The European colonial project relied on epistemic violence to assert dominance, erasing Indigenous knowledge systems and replacing them with Eurocentric epistemologies. AI represents a new frontier of this epistemic hegemony where knowledge production and technological control remain concentrated in the Global North. AI governance, much like international law, is built on the foundations of colonial power structures, as it often reflects the interests and perspectives of dominant global powers, reinforcing existing inequalities in technology access, regulatory frameworks, and ethical standards, disproportionately benefiting wealthy nations while sidelining the needs and voices of developing countries. From the historical imposition of legal norms on the Global South to the present-day concentration of AI development in a few powerful states, the logic remains the same: the production of knowledge and technology is monopolized, while the periphery is relegated to the role of passive consumer.
This is not accidental; it is an extension of a historical project of knowledge suppression. The European colonial project was founded on the suppression of Indigenous knowledge systems, replacing them with Eurocentric perspectives. Today, the dominance of English-language AI models and Western-centric algorithmic assumptions mirrors this epistemic hegemony. From the historical imposition of legal norms on the Global South to the present-day concentration of AI development in a few powerful states, the logic of exclusion, domination, and epistemic violence remains unchanged. The monopolization of AI research, (understood not as the absence of other actors, but as the structural concentration of influence, epistemic authority, and infrastructural control), along with the control of data flows, and the imposition of governance frameworks designed in the Global North, all point to a continuity of imperialism – one that operates through technological means rather than military conquest. We take note of the emergence of innovative tech hubs in cities like Lagos, Nairobi, Johannesburg, Bangalore, Chennai, and Hyderabad, where AI development is thriving independently of Western dominance. These hubs are attracting global partnerships due to their cost-effectiveness and technical expertise. However, it’s important to note that while these hubs are advancing AI capabilities, they often operate within frameworks influenced by global standards and market dynamics that may not fully align with local contexts. Although these tech hubs are in a subaltern space, dominant global structures still reinforce a continuity of imperial logics. These dynamics can perpetuate a form of technological hegemony, where the development and application of AI are shaped by external interests, potentially sidelining Indigenous knowledge systems and local needs. Therefore, the rise of these tech hubs, while promising, does not entirely dismantle the structures of epistemic dominance in AI development.
Furthermore, the dominance of English-language AI models, the Western-centric assumptions embedded in algorithmic decision-making, and the exclusion of Global South voices from the governance discourse mirror the dynamics of colonial rule. AI models, largely trained on datasets reflecting the realities of the global powers, misrepresent and marginalize diverse cultural and linguistic expressions from the periphery. An example is also the use of predictive analytics and biometric tracking, which can reinforce oppressive state power. These technologies echo colonial practices of surveillance, classification, and control over marginalized populations. This epistemic exclusion is not incidental; it is an extension of the historical project of knowledge suppression, where the Global South (with the exception of new economic Global Powers from the South) is rendered an object rather than a subject and agent of technological development. This reality demands a critical interrogation through the lens of TWAIL, which exposes how international legal frameworks, including AI governance regimes, perpetuate existing hierarchies under the guise of neutrality and progress. AI regulatory frameworks further entrench this asymmetry. Major instruments such as the OECD AI Principles and the EU AI Act are drafted in boardrooms in Brussels, Washington, and Beijing (only Global South countries that are economic powers are involved), with little to no input from the Global South. African states, lacking the economic and political leverage to shape the AI agenda, are forced to adopt preexisting norms that do not reflect their sociopolitical and economic realities. This is akin to the legal transplantation of colonial era legal codes, where the periphery is left with rules designed by and for the core. Meanwhile, African states, lacking the economic and political clout to set the agenda, must adapt to rules in whose making they had little say. The discourse around ‘AI ethics’ is similarly shaped by the moral anxieties of the Global North, often ignoring concerns that are more immediate to the Global South, reinforcing colonial era stereotypes, data sovereignty, and the disproportionate impact of automated decision making on marginalised populations.
The discourse on AI ethics, too, is dictated by the moral anxieties of the Global North, often centering on issues such as privacy and transparency while ignoring more immediate concerns for the Global South: the reinforcement of colonial-era stereotypes through algorithmic biases, the erosion of data sovereignty, and the exploitative conditions of AI labour markets. In addition, the environmental concerns associated with AI are indeed pressing, particularly for the Global South, where the intensive energy consumption of AI systems and reliance on data centers that are likely to bring about misinformation, contribute significantly to carbon emissions and water usage, which can exacerbate climate challenges in regions already vulnerable to environmental stressors. Moreover, the extraction of critical minerals for AI infrastructure often leads to ecological degradation and social injustices in these areas. Addressing these issues requires a concerted effort to implement sustainable AI practices and equitable policies that consider the unique challenges faced by the Global South.
The Extractive Political Economy of AI
The extractive political economy of AI governance mirrors the extractive logic of colonial resource exploitation. Data is often described as the ‘new oil.’ The metaphor is apt not only because of its value but also because of its extractive nature. Large technology corporations, predominantly based in the Global North, appropriate data from African populations to refine AI models while offering little in return. This dynamic echo the economic structures of colonialism, where raw materials were extracted from the periphery to fuel industrial development in the metropole. However, some African nations have begun to assert control over their data through legislative measures such as data localization laws. For example, ECOWAS (Economic Community of West African States) has been promoting data protection regulations that emphasize data sovereignty across West Africa. Nigeria has also implemented the Nigeria Data Protection Regulation (NDPR), which mandates that certain data must be stored within the country, ensuring that local populations benefit more directly from the data generated. These legislative efforts are pivotal in challenging the dominance of the Global North over African data and moving toward a more equitable digital economy.
Much like Africa’s natural resources have been siphoned off for centuries, its data is now being harvested to train AI models habitually without consent, compensation, or regulatory oversight. This asymmetry reflects a broader legal architecture where international law facilitates the free flow of resources from the periphery to the core while imposing strict limitations on Global South attempts to assert sovereignty over their own assets. This asymmetry reflects a broader legal architecture where international law facilitates the free flow of resources from the periphery to the core, often through mechanisms like Bilateral Investment Treaties (BITs), which grant multinational corporations favorable conditions for extracting resources. Customary international law and treaties limit the Global South’s ability to control its assets. For example, Bilateral Investment Treaties often protect foreign companies while restricting local regulatory capacity. These agreements prevent countries from passing laws that could negatively affect the profitability of foreign investments, such as data protection laws. Similarly, WTO rules on intellectual property favour Northern corporations, hindering the Global South’s ability to protect its resources or innovate. These legal frameworks reinforce the dependency of the Global South on foreign exploitation.
Facebook’s Free Basics program, which was presented as a way to provide ‘free’ internet to Africans, turned out to be a data-harvesting operation that deepened Africa’s digital dependence on Western platforms. The program exploited weak regulatory frameworks in many African countries, allowing Facebook to collect large amounts of data from users without offering substantial benefits in return. African governments could have acted by creating stronger data protection laws and ensuring that foreign companies followed local regulations. However, international law, particularly agreements such as Bilateral Investment Treaties (BITs), often limits the ability of African governments to regulate foreign companies. These treaties protect foreign investors and prevent local governments from imposing rules that could affect corporate profits. Additionally, international intellectual property laws often favour Northern corporations, making it difficult for African nations to control their digital infrastructure and data. As a result, reliance on data from the Global South not only increases economic dependency but also mirrors the colonial structures of wealth extraction, reinforced by international legal systems that prioritize corporate interests. This leaves African governments with limited power to ensure that the data generated within their borders benefits their people. Yet, in many Global South states, data governance remains an afterthought. Weak regulatory frameworks, underfunded oversight institutions, and a lack of technical expertise create the perfect storm for exploitation.
The legal architecture governing AI reflects a broader trend in international law, where rules are designed to facilitate the free flow of resources, from the periphery to the core, while imposing strict limitations on any attempt by the Global South to assert sovereignty over its own assets. International law often prioritises the interests of multinational corporations and powerful states, creating legal mechanisms that restrict the Global South’s ability to control its digital resources. For example, BITs often grant foreign investors favourable conditions for exploiting local resources, including data, while limiting the ability of host states to impose regulations on foreign companies. Additionally, intellectual property laws, such as those governed by the World Trade Organization, often favour Northern corporations by enforcing monopolies over technology and data. The United Nations Guiding Principles on Business and Human Rights (2011), for instance, while advocating for corporate accountability, remain largely unenforced with regard to AI governance, allowing tech companies to continue operating with minimal oversight in the Global South.
AI Labour Hierarchies and the Global South
The rapid advancement of artificial intelligence (AI) has precipitated a surge in demand for data annotation and content moderation services. Predominantly outsourced to the Global South, these roles are often characterized by low wages, precarious employment conditions, and significant psychological distress. This dynamic mirrors historical patterns of labour exploitation, where the Global North capitalized on the labour of the Global South to sustain its technological and economic dominance.
A poignant example is the experience of content moderators employed by third-party contractors in Kenya. Tasked with reviewing distressing and graphic content to train AI systems for major tech corporations, these workers endure severe psychological trauma. Despite the critical nature of their work, they receive minimal remuneration and inadequate mental health support. This exploitation reproduces colonial labour practices, where extractive economies relied on the subjugation of Indigenous workers to sustain industries controlled by foreign powers.
From a TWAIL perspective, this scenario underscores the perpetuation of colonial power structures within modern AI governance. International labour laws and AI regulatory frameworks, predominantly shaped by the Global North, often fail to address the unique challenges faced by workers in the Global South. This oversight facilitates a system where technological advancements in developed countries are subsidized by the exploitation of labour in less developed regions.
To rectify these imbalances, it is imperative to advocate for the inclusion of Global South voices in the formulation of international AI and labour policies. In the case of Africa, leveraging existing regional frameworks such as the African Union (AU) and the African Continental Free Trade Area (AfCFTA), alongside trade instruments like the Economic Partnership Agreements (EPAs), can be key to advocating for more inclusive policies and ensuring that Africa’s perspectives are integrated into global AI governance discussions. This inclusion would ensure that regulations are attuned to the socio-economic realities of these regions, promoting equitable labour practices. The current AI labour hierarchies reflect a continuation of historical patterns of exploitation, with the Global South bearing the brunt of the burdens associated with technological progress. Thus, there is a pressing need for robust legal frameworks that protect the rights of AI labourers, guaranteeing fair wages, mental health support, and job security. Such measures would not only address the immediate concerns of exploitation but also contribute to dismantling the enduring colonial hierarchies embedded within global AI governance.
Algorithmic Discrimination and the Ghosts of Colonial Law
History has shown that technology is never neutral. From apartheid-era passbooks to the biometric registration of colonial subjects, technological innovations have often been deployed as tools of control. AI is no different. Studies have demonstrated that facial recognition systems exhibit racial biases, with significantly higher error rates for people of African descent. The Global South has also become a testing ground for experimental AI applications with minimal safeguards, treated as mere subalterns in Global AI governance.
In international legal discourse, AI governance reflects longstanding hierarchies of global power. The principle of ‘technological neutrality’– often cited by AI developers- serves as a smokescreen to evade accountability. It echoes the legal doctrines of the colonial period, where international law was weaponized to justify domination under the guise of universality. The failure of international law to intervene meaningfully in AI governance exposes its complicity in sustaining these asymmetries. The 1966 International Covenant on Civil and Political Rights, which guarantees the right to privacy, has been invoked in AI-related cases, but enforcement mechanisms remain weak, particularly in Third World jurisdictions where regulatory capture is prevalent. For example, in countries like Kenya and Nigeria, AI systems are deployed in areas such as policing and healthcare, where ethical and legal protections remain underdeveloped, creating a vulnerability to both privacy violations and exploitation.
Reclaiming AI: A TWAIL Response
A “TWAILian” response to the coloniality of AI demands a radical reconfiguration of existing governance structures. AI is not simply a technological tool; it is a product of historical and political forces. It is also not neutral; it is embedded in power structures that reflect patterns of imperialism. The Global South must reclaim control over AI governance through three key interventions: the recognition of data sovereignty as a principle of international law, the strengthening of intra-African cooperation in AI development, and the adoption of a critical legal approach that challenges the assumptions of technological neutrality.
The fight for resource sovereignty was central to the decolonization struggles of the 20th century. Just as postcolonial states asserted control over their natural resources against the backdrop of economic imperialism, the Global South must now demand control over its digital resources. Data is the new oil, yet it remains largely extracted, processed, and monetized by multinational corporations and governments based in the Global North. African, Latin American, and most Asian states remain relegated to the periphery of this new digital economy, functioning as passive data providers rather than active participants in shaping AI’s future. International legal structures, which often prioritize the interests of powerful Global North nations, exacerbate this marginalization by failing to enforce equitable regulations, while the lack of political will and infrastructure in many Global South countries further hinders their ability to actively participate in AI development and governance. The struggle for control over digital resources must be understood as a continuation of historical fights for economic sovereignty.
The African Union’s Data Policy Framework (2022) represents an important step in this direction, advocating for a rights-based and developmental approach to data governance. However, without robust enforcement mechanisms, the framework risks becoming another well-intentioned but ineffective policy, easily circumvented by corporate actors and state interests from the Global North. The European Union’s General Data Protection Regulation (GDPR) has set a precedent for stringent data protection laws, but the Global South must go further by demanding not only privacy protections but also full control over the economic benefits derived from data.
The African Union’s Continental AI Strategy can be framed as a TWAIL response by arguing that it resists digital neo-imperialism and reclaims Africa’s technological sovereignty. The strategy acknowledges AI as a tool of global inequality, where the Global South remains a passive data provider while the Global North monopolizes AI’s economic benefits. TWAIL critiques this imbalance, positioning Africa’s AI strategy as an assertion of data sovereignty akin to the resource nationalization movements of the decolonization era. Local capacity-building, regional cooperation, and ethical AI grounded in African values challenge the myth of technological neutrality. African values, which prioritize community well-being, respect for local knowledge systems, and a holistic view of human development, offer a framework for designing AI systems that reflect collective, rather than individualistic, interests. In this context, AI systems, often trained on Western datasets, encode biases that marginalize African epistemologies. The AU strategy’s insistence on inclusive, context-sensitive AI development subverts this dynamic, fostering AI that reflects Africa’s diverse socio-political realities. Furthermore, the strategy’s call for South-South AI and intra-African collaboration resists dependency on Northern tech giants. It reimagines AI governance as an emancipatory project, where African states dictate the terms of their digital future. In this sense, the African Union AI Strategy is not merely a policy document, it is an act of defiance against digital coloniality, embodying TWAIL’s demand for equitable technological governance.
South-South cooperation in AI development is critical if the Global South is to escape its position as a technological vassal. AI development today is overwhelmingly dominated by a handful of countries, primarily the United States and China, whose interests shape global AI governance frameworks. Latin America, Africa, and most of Asia cannot afford to remain passive consumers of AI systems that are trained on biased datasets and optimized for interests that do not align with their developmental priorities.
India’s National AI Strategy, which emphasizes ethical and locally relevant AI, offers a potential model for Global South states. It prioritizes AI applications that address social challenges such as healthcare, agriculture, and education sectors often neglected by profit-driven AI development in the Global North. African, Latin American and Southeast Asian nations have already started establishing collaborative AI research hubs, such as the Latin American Artificial Intelligence Network (REDIA) and the ASEAN Smart Cities Network, which promote cross-regional cooperation. Additionally, joint funding initiatives like the Singapore-India AI Cooperation have been instrumental in pooling resources for the development of AI technologies tailored to the unique challenges and opportunities of these regions.
Moreover, the current structure of AI development reinforces technological dependency, where Global South nations are forced to adopt systems designed without their input. TWAIL scholars argue that genuine independence in AI requires resisting this dependency through strategic investments in local AI ecosystems and cross-regional partnerships that counteract the monopolization of AI expertise by a few dominant powers. A fundamental assumption embedded in mainstream AI governance is that technology is neutral and universally beneficial. For instance, facial recognition systems, often trained on datasets predominantly featuring lighter skin tones from the Global North, exhibit significantly higher error rates for individuals with darker skin tones prevalent in many Global South populations, thereby perpetuating racial discrimination in applications such as surveillance and policing. International legal frameworks must reflect this reality to avoid perpetuating the illusion of objectivity free from ideological and economic interests.
The United Nations has recently called for an international AI treaty, presenting an opportunity to insert TWAIL perspectives into global policymaking. Such a treaty must not simply replicate existing power hierarchies but instead actively redistribute authority over AI governance. The current AI governance discourse largely mirrors colonial-era global governance structures, where the Global South is relegated to the role of passive rule-taker rather than rule-maker. The UN AI treaty should incorporate provisions ensuring equitable representation of Global South voices in AI standard-setting bodies.
Additionally,, international law must reckon with the ways AI is being deployed to entrench existing inequalities. The use of AI-powered surveillance in Africa and Asia, often under the guise of security enhancement, disproportionately targets marginalized communities while serving the interests of both domestic elites and foreign powers. TWAIL scholars must expose and resist these new forms of digital imperialism, arguing for a human rights-centered and sovereignty-driven approach to AI regulation. Reclaiming AI from its colonial entanglements requires nothing less than a radical transformation of global governance structures. Data sovereignty must be enshrined as a fundamental principle of international law, South-South AI cooperation must be prioritized, and the myth of technological neutrality must be dismantled. Without these interventions, AI will remain an instrument of imperial domination rather than a tool of liberation.
Concluding Thoughts: Resisting AI Imperialism
The trajectory of “TWAILing” AI governance hinges on whether the Global South remains constrained by the passive role imposed upon it or asserts agency to disrupt and transform that imposed order. AI governance is not merely a technical issue but a legal and political struggle, one that echoes the broader contestations of international law. TWAIL reminds us that law is never neutral; it is a product of power shaped by historical struggles and contemporary resistance. If AI is to serve the interests of all humanity rather than a privileged few, then the Global South must reclaim its agency in shaping this future.
Whether this future is one of digital servitude or technological self-determination will ultimately depend on the ability of Third World states and scholars to reclaim the narrative and to shape the law accordingly. The future of AI governance is not predetermined; it is a site of contestation shaped by the political will of the Global South to resist its continued marginalization.
Jake Okechukwu Effoduh is an Assistant Professor of Technology Law at the Lincoln Alexander School of Law of Toronto Metropolitan University, Canada. Miracle Okumu Mudeyi is currently pursuing the Advocates Training Program at the Kenya School of Law. His research interests include sovereign debt, Public International Law, Constitutional Law, Third World Approaches to International Law, AI governance, and human rights.
