Greening Extractivism: Justifying AI supply chains in Canada


Kailey Walker, Queen's University

Drawing on environmental media studies, theories of data colonialism, and Povinelli’s (2016) geontologies, this paper analyzes the discursive strategies deployed to legitimize AI extractivism in Canada. In the face of anthropogenic climate change, the Big Tech industry broadcasts its embrace of environmental standards with carbon-neutrality, re-forestation, and water recycling programs. Meanwhile, the industry’s supply chains are responsible for disproportionately harming Indigenous communities in the context of resource extraction, generating pollution and toxic waste to manufacture microchips and batteries, and exhausting vast reservoirs of natural resources to operate data centers. Emerging literature on this debate thus situates AI as an extractive industry, guilty of exploiting humans and non-humans alike for profit and power. To justify the socio-environmental costs of these extractive sectors, discursive strategies of social licencing are adopted by corporations to green extractivism: AI is conflated with climate solutions to not only distract from (i.e., greenwashing) but rationalize the devastating socio-environmental effects of digital supply chains. At a time when Canada is making substantial investments into both AI and ‘green’ mineral industries, discursive strategies are similarly deployed by the state to frame the harms of resource extraction and expenditure as necessary for green energy transitions. The aim of this paper is to clarify the colonial logics that underpin green extractivism in Canada. Three main theoretical resources support this research: 1) environmental media studies; 2) theories of digital and data colonialism, and 3) Povinelli’s (2016) geontopower. This interdisciplinary scholarship studies the constitutive roles that tech corporations play in the composition of Big Data Ecologies, how digital supply chains amplify historical forms of colonization through complex arrangements of practice, materiality, and discourse, and how extractivism relies on colonial ontologies which distinguish ‘humans’ from what is ‘natural’ or ‘less than’ human. Taken together, these debates emphasize how power and domination in resource governance result in uneven outcomes along the lines of historical colonialism, ongoing settler colonialism, and environmental racism. Drawing on decolonial and qualitative methodologies, this paper puts forward a critical discourse analysis of dominant narratives of ‘green AI’ and ‘tech minerals’ as advanced by 1) the state in strategies, policies, and public statements, and 2) the ten largest tech corporations in Canada through their environmental reports, press releases, etc. Narratives of ‘land as resource’, ‘untapped mineral potential’, and the necessity of AI for ‘green futures’ are identified between these sources. I argue that justifications of AI extractivism significantly depend on colonial ontologies that devalue the natural world as Nonlife, which highlights the Western epistemological frameworks that enable and sustain the harms of AI supply chains. This paper contributes to the burgeoning debate on AI and inequality by situating extractivism as part of enduring settler colonialism in Canada. In doing so, this analysis adds specificity to theories of data (or digital) colonialism by clarifying how AI extractivism depends on the ongoing suppression of Indigenous knowledge about Land and Life – while simultaneously demarcating what (and who) is expendable in the quest for green futures.

This paper will be presented at the following session: