Creating certainty where there is none: Artificial intelligence as political concept

Recent developments in research on Artificial Intelligence (AI) have prompted a growing politicisation of AI. In this paper, we critically analyse how AI is being construed in public discourse and with what political implications. Approaching AI as a mobilising political concept, and focusing on the public pronouncements made by influential tech commentators, we identify and subject to technical and critical scrutiny four key themes in contemporary discourses on AI. First, we show how AI discourses endorse anthropological commitments that create false equivalences between human and artificial intelligence, and suggest that all are equally affected by AI. Second, we demonstrate that AI discourses unproblematically indulge in agential constructs, which ascribe agency to AI while obfuscating the role of humans in its development. Third, we explain how the economic assumptions made by these discourses support specific political interests. Finally, we show that discourses on AI endorse a set of temporal assumptions that reduce the space for democratic intervention. We conclude that AI is becoming more than what its ‘technical’ specifications would warrant; however, this is happening in a way that limits the space for democratic engagement with, and control of, the technology itself.