AI systems encode power: How algorithmic maps exclude marginalised knowledge and reshape societal reality
Original framing: “Why AI output is always a map, never a territory” — bing news
Indigenous data sovereignty movements (e.g., Māori data governance in Aotearoa), historical parallels in cartography and colonialism (e.g., how maps justified land theft), structural critiques of surveillance capitalism, and the role of Global South labour in training AI systems. The original framing also omits the epistemic violence of reducing complex social realities to quantifiable datasets, as well as non-Western epistemologies like Ubuntu philosophy or Andean relational ontologies that centre collective well-being over individual data points.
Medium structural omission detected in mainstream coverage.
The narrative is produced by tech industry-aligned media (e.g., The Mandarin) and Western academic institutions, serving corporate interests in legitimising AI deployment while obscuring its extractive foundations. The framing serves Silicon Valley’s 'move fast and break things' ethos, positioning AI as inevitable while ignoring its role in consolidating epistemic power in the hands of a few. It obscures the role of venture capital, surveillance capitalism, and neoliberal governance in shaping AI systems, which disproportionately harm marginalised communities through biased training data and opaque decision-making.
Marginalised communities—particularly indigenous peoples, Global South populations, and gig workers—are disproportionately affected by AI’s 'map-territory' problem. Their knowledge is mined for training data but excluded from system design, as seen in cases like facial recognition bias against darker-skinned people. Grassroots movements like *Data for Black Lives* and *Algorithmic Justice League* demand accountability, but their voices are often sidelined in mainstream tech discourse. Centring these perspectives reveals AI as a tool of structural violence, where 'maps' justify exclusion and exploitation.
AI’s 'map-territory' problem is not a technical glitch but a manifestation of deeper epistemic violence, where algorithmic systems encode colonial, capitalist, and patriarchal logics.