AI in the age of technoscience: on the rise of data-driven AI and its epistem-ontological foundations

The chapter offers a critical discussion of the epistemological and ontological foundations of AI. It begins with a short history of AI focused on the epistemological and ontological premises foundational to the three AI approaches that have dominated its historical development, namely, symbolic, connectionist, and data-driven AI. It then presents some early and more recent critiques of AI technologies that are informed by science and technology studies, including more recent developments in algorithm and critical data studies. The chapter is to contribute toward a reflection on the all-too-often implicit assumptions entailed by AI per se and by its new instantiations, thus laying the analytical groundwork for shaping alternative AI technologies in the future. Throughout the last few years, reports of racist risk assessment tools employed in the US criminal justice system, of sexist recruiting tools, or of highly stereotypical digital assistants have highlighted the fact that AI programs are far from being neutral and objective.