COLONIAL MAPPING OF ADVANCE AUTOMATION: EAST INDIA COMPANY AS AI

Abstract

This paper argues that AI systems must be understood as extensions of centuries-old platform infrastructures rooted in empire. A telling anecdote comes from a senior Microsoft scientist who asked a generative model to depict “an Indian girl under a tree reading a book.” Instead of producing her desired childhood reflection, the system defaulted to white, Western images—revealing how foundation models encode colonial value systems. Like Stuart Hall’s encoding/decoding model, the circulation of meaning is pre-structured by Western archives that dominate training data. This dynamic is not new. The British East India Company operated as the first global platform: it centralized trade through charters, maps, tariffs, and racial ranking systems, sustaining a multi-sided market without producing its own content. Just as the EIC tweaked terms and conditions to regulate life across continents, modern AI platforms hard-code racial capitalism into digital infrastructures. Contemporary Platform Studies often miss this continuity, framing platforms as neoliberal novelties rather than as inheritors of colonial techniques of privatization, dispossession, and labor precarity. Today, national initiatives such as India’s foundational AI model demonstrate both possibility and peril. While framed as sovereign alternatives to Western dominance, such projects risk reproducing exclusion by privileging some languages and identities over others. In contrast, Indigenous data sovereignty movements—such as Te Mana Raraunga in Aotearoa—foreground communal ownership and resistance to state or market capture. This paper situates “Global Southing” not as a guarantee of decolonization but as a contested terrain where platform power may either reinforce or unsettle colonial continuities.