Jörg Volbers

FU Berlin

Our Technological Form of Life. Wittgensteinian Lessons on the Moral Dimensions of AI

28 February 2023, 17h00 (Lisbon Time — GMT+0)

Sala Mattos Romão (Room C201.J – Department of Philosophy) | School of Arts and Humanities – University of Lisbon

Abstract

Technologies with so-called Artificial Intelligence (AI) are a world-wide reality and do already deeply pervade daily life. Yet in contrast to other technologies, AI is employed for a quite distinct type of tasks, such as recognizing faces or diagnosing diseases, all of which involve some kind of judgment. Passing these tasks to machines raises a special kind of moral problem: How can we rely on such an AI-guided process, given that this guidance is made by an automaton which cannot be made accountable for what it decides? In response to that problem, it is often suggested to refer to regulations. A burgeoning industry of AI ethics spends its time devising rules, principles, or ethical frameworks, to which an ethically well-behaved AI should defer. I will argue, however, that we should see this problem as an instance of the so-called “problem of other minds”: The ’reasoning’ of AI machines is constitutively incomprehensible to us (“black box”), and yet we are forced to interact with them. In this view, it is a mistake to believe that ethical rules could somehow turn AI into ethically responsible machines. For this reason, we should be wary of the current attempts to treat AI as a problem of regulation only, nor can we ever hope to solve it by gaining a better epistemical insight into the inner workings of artifi cial mind (so-called “explainable AI”). Rather, it manifests the challenge to understand ethics, and morality, as a practice, or a “form of life,” as Wittgenstein calls it.