Decoloniality and AI

Technology and geopolitics are inextricably linked, reshaping the global balance of power and existing inequalities. It suggests that technological breakthroughs also shape the socio-political and economic realities of the Global South in ways that are reminiscent of its history with colonialism and its persistent contest with coloniality. In this regard, the unprecedented launch and mainstreaming of Artificial Intelligence (AI) is an apt case in point because its revolutionary prospects for the way societies function and human beings interact co-exist with imperial dynamics [i] that are paving the way for new forms of exploitation, subjugation, and oppression.
Moreover, since the progress in this technology is ongoing and unavoidable, with UN Trade and Development (UNCTAD) projecting [ii] a $4.8 trillion increase in the AI industry by 2033, integrating decolonial perspectives with the advancements in and implementation of AI becomes imperative in ensuring that this period of transition yields more equitable outcomes.

How is AI redefining colonialism?
The existing process of resource exploitation and unequal social relations that fortify appropriation [iii] in the development and deployment of AI renders it a new form of colonialism. Numerous studies have made this argument using terms such as “techno-neocolonialism,” “data colonialism,” “digital apartheid,” and “AI colonialism” in which the extractivist nature of this technology emerges as the common denominator of its colonial nature. For instance, the physical infrastructure, including the massive data centers and excessive energy-consuming supercomputers required in running and speeding the AI boom, needs vast amounts of land to build on and huge quantities of water to cool them down [iv] which results in massive environmental exploitation. Similarly, the miners and factory workers who work in hazardous conditions to ensure smooth “data crunching” bear the implicit brunt of running the AI machine whereas the “data annotators” remain invisible and underpaid as they test the final AI models to deliver on the promise of unprecedented innovation and prosperity. Kenyan employees responsible for training AI models as moderators describe [v] their work not only as “low pay” but also “monotonous” and “traumatizing.” Since the latter form of exploitation against the environment and labor occurs predominantly at the discretion and dictation of powerful players in the Global North, the colonial element manifests itself in a form of necroexportation [vi] that systematically transfers the “detrimental impacts” of AI to the Global South.

Why does the Global South need to decolonize AI?
While physical extractivism replicates the colonial model of land grabs and labor exploitation, the marriage between data colonialism and data capitalism [vii] threatens to upload the historical continuity and structures of coloniality in the Global South. AI models codify the lived experiences of Global South citizens into data and reduce them to a commodified entity that can be exploited and capitalized upon- like the “colonized and objectified man” in Fanon’s view. User data is extracted [viii] from developing nations without ethical considerations for algorithmic training of technologies that are only serving the consumers and investors in the Global North, evident in the presence and services of Big Tech in South Africa to harness the power of users’ data for profit [ix]. On the other hand, algorithmic decision-making tools trained on nonindigenous and non-localized data, become tools of surveillance, policing, and patrolling the colonized bodies of the Global South, reducing it to a testing ground for politically controversial, ethically ambiguous, and increasingly dangerous AI-powered technologies. The rapid creation of a centralized, coordinated, entirely privatized mass surveillance operation [x] reviving digital apartheid in South Africa is an apt example. A more contemporary and perhaps more suitable example comes from the Gaza Strip where Operation Lavender [xi] used “surveillance data to rate people based on their suspected likelihood of association with a militant group,” exacerbating the indiscriminate loss of human life and scale of the ongoing genocide. Such forms of extractivism directly contributing to exploitation and oppression of a disempowered global majority underpins the need for decolonial solutions moving forward.

How to decolonize AI?
Decolonizing AI requires identifying the colonizer as the new empire. Referring to Karen Hao’s framing of AI companies, particularly, OpenAI, and others in the Silicon Valley, as individual empires provides a framework to understand how corporate greed, market-power competition, and profit motive drives the entire AI industry to its colonial pursuits and practices. Recognizing that not only do these “empires” not benefit the Global South, but promise progress for only a few in the Global North at the expense of their resources, land, people, and labor is the first step to challenging their impunity and authority. Addressing the false universalism [xii] rooted in Western values as well as the racialized elements of control and subjugation embedded within these models is the natural next course of action. Grassroots level responses may contribute to decolonizing AI, as evident in the following two cases. Ameera Kawash [xiii], while recognizing the limited scope for decolonizing AI with its quintessential need for exploitative scalability, suggests building more curated data training models that prevent resulting biases. Furthermore, concerned with the collection of their cultural and linguistic practices for training AI models for the government, New Zealand’s indigenous Maori community is not only building their AI models to preserve their Te Reo Māori language but also advocate [xiv] against OpenAI’s (mis)use of their data for ‘Whisper’ software. Finally, organized resistance in the name of redistribution of power, as evident in Chile where indigenous communities are protesting [xv] the scale of mining by “flying black flags,” “physically block the roads,” and “contracting lawyers.” Deep-rooted systems of exploitation require strategic responses from a collective front which must be disruptive enough to be a form of decolonial resistance.

Conclusion
The disproportionate concentration of power over AI technologies and the benefits that result from its innovation and progress in the hands of corporate players based in the Global North adds a colonial dimension to AI. It requires a people-driven decolonial reimagination of the ways in which resources are extracted, data is collected, AI models are trained, and consequently, deployed and distributed, which will in turn ensure that the Global South is no longer a victim of exploitative digital colonialism but an active beneficiary of an AI-powered
revolution.

[i] Tech Policy Press. 2025. “Decolonizing the Future: Karen Hao on Resisting the Empire of AI | TechPolicy.Press.” Tech Policy Press. https://www.techpolicy.press/decolonizing-the-future-karen-hao-on-resisting-the-empire-of-ai/.
[ii] UN Trade and Development (UNCTAD). 2025. “AI market projected to hit $4.8 trillion by 2033, emerging as
dominant frontier technology.” UNCTAD. https://unctad.org/news/ai-market-projected-hit-48-trillion-2033
emerging-dominant-frontier-technology.
[iii] Couldry, Nick, and Ulises A. Mejias. 2019. The Costs of Connection How Data Is Colonizing Human Life and
Appropriating It for Capitalism. N.p.: Stanford University Press. https://doi.org/10.1515/9781503609754.
[iv] Wray, Ben. 2025. “Sam Altman’s AI Empire Relies on Brutal Labor Exploitation.” Jacobin.
https://jacobin.com/2025/07/altman-openai-artificial-intelligence-labor-environment-deepseek.

[v] “’It’s destroyed me completely’: Kenyan moderators decry toll of training of AI models.” 2023. The Guardian.
https://www.theguardian.com/technology/2023/aug/02/ai-chatbot-training-human-toll-content-moderator-metaopenai.
[vi] Regilme, Salvador S. 2024. “Artificial Intelligence Colonialism: Environmental Damage, Labor Exploitation, and Human Rights Crises in the Global South.” SAIS Review of International Affairs 44 (2): 75 – 92.
https://dx.doi.org/10.1353/sais.2024.a950958.
[vii] Mohamed, Shakir, Marie-Therese Png, and William Isaac. 2020. “Decolonial AI: Decolonial Theory as
Sociotechnical Foresight in Artificial Intelligence.” Philosophy & Technology 33:659 – 684.
https://doi.org/10.1007/s13347-020-00405-8.
[viii] KPONYO, Jerry J., Dickson M. FOSU, and Frederica E. OWUSU. 2024. “Techno-neocolonialism: an emerging risk in the artificial intelligence revolution.” Trayectorias Humanas Trascontinentales 18.  https://doi.org/10.25965/trahs.6382.
[ix] Muldoon, J., Wu, B.A. Artificial Intelligence in the Colonial Matrix of Power. Philos. Technol. 36, 80 (2023).
https://doi.org/10.1007/s13347-023-00687-8
[x] Hao, Karen, and Heidi Swart. 2022. “South Africa’s private surveillance machine is fueling a digital apartheid.” MIT Technology Review. https://www.technologyreview.com/2022/04/19/1049996/south-africa-ai-surveillance-digital-apartheid/.
[xi] “Questions and Answers: Israeli Military’s Use of Digital Tools in Gaza.” 2024. Human Rights Watch.
https://www.hrw.org/news/2024/09/10/questions-and-answers-israeli-militarys-use-digital-tools-gaza

[xii] Mollema, Warmhold J. 2024. “Decolonial AI as Disenclosure.” Open Journal of Social Sciences 12:574-603.
https://doi.org/10.4236/jss.2024.122032.
[xiii] UntoldMag. 2025. “What does decolonising AI really mean? An interview with artist Ameera Kawash.” Global Voices. https://globalvoices.org/2025/04/18/what-does-decolonising-ai-really-mean-an-interview-with-artist-ameera-kawash/.
[xiv] Grimshaw Club. 2025. “The use of Māori knowledge in AI: How does it affect indigenous data sovereignty and human rights?” https://www.grimshawclub.org/post/the-use-of-m%C4%81ori-knowledge-in-ai-how-does-it-affect-indigenous-data-sovereignty-and-human-rights.
[xv] Hao, Karen, Varsha Bansal, and Divsha Bhat. 2025. “Indigenous Chileans resist AI-fueled copper and lithium
mining.” Rest of World. https://restofworld.org/2025/ai-resource-extraction-chile-indigenous-communities/.