Decolonising Digital Futures

On May 9th 2024, the EthicsLab hosted a webinar featuring Sebastián Lehuedé, who critically explored the colonial dimensions of the current global digital order and whether embracing digital sovereignty is a suitable response.

Lehuedé employed a decolonial lens to interrogate the ethical dimensions of digital and data-driven technologies like Artificial Intelligence, particularly considering its environmental impact. Central to this was the concept of ‘coloniality’ from the Latin American modernity/coloniality research program, which recognises the persistence of colonial power dynamics, identities, and inequalities well after independence and the end of direct colonial rule. Consequently, Lehuedé explored how AI and data technologies reproduce or entrench structures of colonialism. 

Lehuedé focused on the political economy and environmental dimensions of what Nick Couldry and Ulises Meijas have coined ‘data colonialism’. The political economy of data colonialism centres on relations in the production and exchange of data and the extractivism inherent within the datafication and commodification of human life by large corporations and governments. This exploitation also extends to the extraction of minerals and other natural resources, their production and exchange mediated by colonial institutions and divisions of labour. 

Environmental colonialism captures the outsourcing of ecological damage. While profits are concentrated in countries like the US, the environmental impact of AI is dispersed across the majority world through the creation of data infrastructures and mineral extractivism. AI climate solutionism – or the presentation of AI as a technology uniquely capable of solving the climate crisis – is another subtle but equally concerning discursive dimension to environmental colonialism. Not only does such a presentation obscure the ecological impact of digital technologies and data infrastructures, it also marginalises alternative non-technical ways of relating to nature by indigenous communities and knowledge systems, for example.

These insights on the AI-colonialism-environment nexus were situated in Chile, where Lehuedé conducts empirical research. Over the last 15 years, a devastating drought has meant that the construction of data centres in Santiago and lithium extraction in the Atacama desert, propelled by the AI boom, have taken on added significance around access to water and energy, prompting emergent networks of solidarity and resistance among local and indigenous communities. 

Lehuedé considered the merits of embracing digital sovereignty frameworks as an alternative technological imaginary in response to the coloniality of AI and data-driven technologies. He argued that data sovereignty frameworks are attractive because they potentially foster a polycentric world order by challenging US digital hegemony. However, they are still limited in some regards, entrenching economic dependency and complicity with capitalist extractivism and environmental damage (Lehuedé 2024).

The ensuing discussion registered several key insights:

1. Materialist conception of AI. The ontology of AI, with talk of data, clouds, and networks, can often feel ethereal. Much of this webinar was concrete, foregrounding physical infrastructures such as servers, data centres, minerals, water, land, people, and the environment as the material bedrock of various political, economic, and social relations.

2. Environmental stakes. Scholars and activists' work foregrounding the ecological impact of AI infrastructures and algorithms is significant because such accounts have been largely missing in public discourses on AI. In AI ethics, for example, more traditional concerns have included privacy, copyright, bias, discrimination, and a spuriously looming robot uprising. Lehuedé underscored the imperative of integrating environmental considerations into AI ethics discussions, asserting the urgency of acknowledging and mitigating the ecological footprint of digital technologies.

3. The politics of voice and vulnerability. Lehuedé reflected on a quote from an indigenous leader asserting the threat of extinction facing his people due to lithium mining and its intensification of water scarcity. As he astutely observes, this is critically and qualitatively different from prevailing discourses which forebode humanity’s extinction with the rise of super-intelligent machines. Vulnerability, like ethics, is political. What truly counts is rarely counted, and what is often counted perhaps shouldn’t count. 

4. International networks of solidarity. Lehuedé also noted the potential for emergent networks of solidarity and resistance based on shared experiences and conceptual resources. He argues that regions such as Africa and Latin America hold relatively similar positions within the value chain and infrastructure of digital technologies. From these shared vantage points, cases of resistance that incorporate not only communities affected by data centres and extractivism but also the data workers and other actors involved in the value chain of artificial intelligence could be connected. There are also shared conceptual resources, for example, in the tradition of relational ontology. Some activist protest repertoires discussed in the webinar were striking in recognising the agency of nature (water in this case). This undergirds alternative ways of relating to nature that democratise moral citizenship beyond the binary between humans and non-humans.