Generic link - The expansion of IPR leads to the embracement of digital colonialist ideas

Quijano (2007) conceptualised the colonial matrix of power (“patron de poder colonial”) to describe the current world-systems entanglement with historical colonialism. The colonial matrix of power is an organising principle involving exploitation and domination across four interrelated domains: “control of economy (land appropriation, exploitation of labour, control of natural resources); control of authority (institution, army); control of gender and sexuality (family, education) and control of subjectivity and knowledge (epistemology, education and formation of subjectivity).” (Mignolo, 2007, 156). Although we draw on all four domains in our analysis, the article focuses on the control of economy (namely labour and resources) and the control of subjectivity and knowledge. Quijano (1993) argues that within the matrix, these domains are naturalised around the colonial articulations of “objective” and “scientifc” notions of race and racial classifcations. These categories underpin the racial division of labour, resources and capital that constitute one of the most visible organising principles structuring modern societies.

The exercise of economic dominion over both labour and resources is closely intertwined with the ideological construct of European technological preeminence used to justify European colonialism. European colonisers imposed authority in science, culture and societal administration while communities from Africa and Asia were expected to acculturate and obey while providing labour to the colonial project (Adas, 1989). Quijano (2016) links the manipulation and control of technological resources by colonial powers to the analytic of modernity/coloniality. He argues that technologies of communication and transport were used to instrumentalise the superiority of European modernity, which were mutually constitutive with the violence of colonialism. This highlights the important fact that the colonial legacy of Western dominance, in terms of both modernity and technological progress, continues the legacy of colonial domination despite the elimination of political colonialism in many areas of the world (Quijano, 2007).

The modernity/coloniality analytic emphasises that the imaginaries and epistemologies of AI development are inextricably predicated upon rationality and universality and are not divorced from coloniality. Indeed scholars of sociotechnical imaginaries and feminist science and technology studies remind us that technology is co-created by the social, cultural and political environment around it (Jasanof, 2004). In the case of algorithmic technologies, these are modelled on visions of the social world, with their outcomes infuenced by the agendas of corporate and political elites who stand to gain from the economic and political changes engendered by the deployment of these technologies (Beer, 2017).

The modernity/coloniality analytic seeks to critically examine the logics of colonialism with respect to both the economy, but also knowledge and cultural production. This was closely tied to the hierarchical paradigm of European rationality as a universal form of knowledge – rendering Europeans as rational “subjects” while others were merely “objects of knowledge of domination practices” (Quijano, 2007, 174). This served the epistemological authority for the destruction of language and cultural practices during colonial rule. The epistemologies of technology are also infuenced by the colonial matrix of power; local technologies were viewed as primitive and traditional (Quijano, 2007). The continuity of this logic can be traced to the post-war discourses about international development, led by Western institutions such as the World Bank, that infantilised non-Western and local knowledge and technology in favour of unilaterally imposing Western development practices (Escobar, 2007).

(Escobar, 2007). This dynamic is also evidenced in the epistemology of academic disciplines such as anthropology that have perpetuated the subjugation of non-Western knowledge and culture, long after former colonies gained independence. We argue that these logics can be traced throughout the development and deployment of AI technologies, from Big Tech’s concentrated ownership of AI’s intellectual property and computing infrastructures, to the export of technologies and their embedded values into the majority world. It is through these exercises of power that a Western mode of knowing has been constituted and reinforced as the dominant imaginary about AI which is tightly bound to a story of modern growth and progress. However, as the colonial matrix of power analytic elucidates, the vestiges of colonialism persist and necessitate a de-linking from narrow ways of thinking about colonial relations in order to accomplish the unfnished and incomplete dream of decolonisation

Emerging scholarship from the social sciences, across platform studies, critical data studies and media studies, trace the legacies of colonialism in digital technologies and their data infrastructures. These can broadly be parsed into digital colonialism, data colonialism, and decolonial approaches to technology and AI development. The study of “digital colonialism” emerged as a theoretical lens for analysing how the fow of data, technology and value, usually from wealthy Western economies to the Page 5 of 24 80 J. Muldoon, B. A. Wu 1 3 majority world, reinforce colonial legacies and new forms of economic and political domination. Theorists trace how technology companies exercise new forms of control over the majority world through manufactured dependence on their digital tools and platforms, subjecting populations to new forms of surveillance and data extraction (Avila, 2020; Casati, 2013; Coleman, 2018; Kwet, 2019b). Avila (2018, 2020) shows that major technology companies in the US seek to obtain control over critical infrastructure in order to gain access to data produced by the majority world. She defines digital colonialism as “the deployment of imperial power over a vast number of people, which takes the form of rules, designs, languages, cultures and belief systems serving the interests of dominant powers” (Avila, 2020, 1). Avila (2020) asserts it is not states but “technology empires” that hold the dominant positions through their control of “digital infrastructures, data and the ownership of computational power” in addition to solidifying their power through favourable global trade agreements.

Kwet (2019b) has described digital colonialism as a “structural form of domination” which technology companies, and particularly Big Tech, exercise through “three core pillars of the digital ecosystem: software, hardware, and network connectivity.” Kwet argues this is a neo-colonial form of “imperial control,” particularly due to the integration of the surveillance apparatuses of corporations into the intelligence services of the US state and the dominance of a Silicon Valley ideology across the globe. Kwet (2019a) highlights the relationship between Government elites and Big Tech in South Africa, citing the example of Operation Phakisa Education, a failed education reform attempt from the South African government aiming to “leverage ICTs to improve basic education.” He argues that Big Tech investment in majority world classrooms, through programs like Microsoft Partners in Learning and Google Classroom, create infrastructure dependencies amongst youth while biassing local tech ecosystems towards Big Tech. This is also evidenced through Big Tech’s provision of free internet via platforms like Facebook which centralise Facebook’s control over communications and news, and drains local advertising revenues. In both examples, the historical mechanisms of surveillance capitalism are perpetuated, extending the use of colonial patterns of racialised extraction and domination through to the ongoing policing of Black bodies. Abeba (2020) illustrates how the ‘algorithmic colonisation of Africa’ imposes Western values onto locales while impoverishing local tech development, citing that Nigeria imports 90% of all software used in the country. In a similar vein, but from a legal perspective, Coleman (2018) analyses how data protection laws and regulations are often absent in African countries which has given rise to a modern day “scramble for Africa” in which tech companies ofer purportedly “free” and altruistic services in order to harness the power of users’ data for profit. In each of these cases, the authors argue that issues of power and technology cannot be discussed without attention to local context, history and the values embedded in technologies. This research agenda is wide reaching and lays the groundwork for this article to further systematise these ideas and to integrate a political economy perspective of the production of AI.