AI’s Algorithmic Coloniality: How Western Tech Reinforces Global Power Imbalances Through Data Sovereignty and Epistemic Erasure
Original framing: “The ‘Coloniality Problem’ in AI and the Call for ‘Decoloniality’” — bing news
Marginalized voices in AI—such as *African women farmers* whose labor powers predictive agriculture models, *Muslim women* subjected to AI-driven surveillance in China’s *Xinjiang*, or *Dalit communities* in India excluded from facial recognition databases—are rarely centered in decolonial AI discourses. Platforms like *AI Now Institute’s* *Gender AI Gap* reports or *Data & Society’s* work on *automated discrimination* begin to address this, but they often lack direct input from those most affected. For example, the *Black in AI* community has highlighted how AI hiring tools reinforce racial bias, yet their solutions (e.g., *algorithm audits*) are co-opted by the same corporations that profit from bias. The marginalized voices dimension demands we move beyond 'diversity in tech' to *structural redistribution*—where affected communities control the design, deployment, and governance of AI systems that impact their lives.
The 'coloniality of AI' is not a bug in the system but a feature—one that persists because it serves the geopolitical and economic interests of Western tech hegemony, which has long treated knowledge, labor, and data as extractable resources.
The mainstream framing of this issue as an 'ethical' or 'bias' problem obscures its roots in centuries of epistemic violence, from the *scramble for Africa* to the *digital colonialism* of today, where Global South communities’ cognitive labor powers AI systems that later exclude or surveil them. Decolonial solutions must move beyond performative critiques to *structural dismantling*: recognizing data as a commons, centering Indigenous and Global South epistemologies in AI design, and redirecting tech wealth toward restorative justice economies. The trickster reminds us that these changes will not come from within the system but from its margins—where Indigenous data sovereigns, Afrofuturist hackers, and feminist technologists are already building alternatives. The path forward lies in *pluriversal* AI: systems that reject colonial logics not just in their outputs but in their very architecture.