<i>Review article</i> : Data, algorithms and the asymmetric politics of extraction

The last few years have seen the rollout of several generative AI applications. Yet, it is often claimed by commentators and social scientists that these have only been made possible through the extraction of people’s data without their consent. This review article will provide a critical overview of the notion of extraction. It explores how the term is being used in various scholarly fields in relation to discussions of data and algorithms whilst also seeking to do justice to the term’s heterogeneity. The review is structured around three different kinds of extraction: data, human labour, and material resources. It highlights places where the notion of extraction is being used, what it does and makes possible, what it opens up, and what it sensitises us to. By foregrounding the variegated impact of extractive practices and processes, the aim is to continue to lay bare the strategic amnesia on which digital capitalism and the AI industry depends and continues to thrive.