Computing Human Oversaight: Law/Apparatusvision/Agency

Computing Human Overs(a)ight is a conceptual exploration of human responsibility for oversight in high-risk AI systems, as introduced in the European Artificial Intelligence Act (Art. 14). Investigating the legal and computational framework of the high-risk systems, and human responsibility within, we amplify possible enactment and embodiment of this article — we extract it as a phenomenon from the law so to understand the political and ideological notion of agency in the automated systems. The accelerated dissemination of the tools, models and products under the roof term AI in the scope of the last decade, has led us to their legitimacy in the eyes of the law, even in high-risk operations. Article 14 invokes the setting for the problematization of current intertwinements between computational and legal acts and abstractions. Such an approach can help us to understand the societal consequences of human-machine operations, addressing the depositions between computation and human agency, transparency, responsibility, and dignity. Relying on critical media studies that address the computational processes through a distinction between operational and representational notions of a computational image, we are questioning the factuality of what is seen and what can be overlooked in the human oversight of high-risk AI systems.