Wednesday 13 May, 2021
Title: From Responsibility Gaps to Responsibility Maps
Abstract: Recent work in moral philosophy has been concerned with issues of responsibility as they relate to the development, use, and impact of artificial intelligence (AI) systems. Some have argued that the emergence of increasingly agential and sophisticated AI-driven systems will create situations where we demand ascriptions of responsibility, but no fitting candidate can be found. In normal situations, the manufacturer or operator is held responsible for the consequences that follow from the operation of their machine. Machines that learn, due to their behaviour being in principle unpredictable from the perspective of their creators, might therefore pose a unique challenge to our responsibility practices, creating a responsibility-gap. Due to the pluralism of responsibility as a concept, there is no single responsibility-gap. Rather, there may be many such gaps, each with their own unique conditions of emergence. My focus will be on the normative senses of responsibility, and whether in these cases a responsibility-gap emerges due to AI. Ultimately, I argue that there is no AI created responsibility-gap. However, I maintain that reflecting on the ways AI and technology challenge our responsibility practices is both practically necessary and philosophically fruitful.
Speaker: Fabio Tollon is currently a doctoral candidate at Bielefeld University, funded by the DFG research training group 2073 “Integrating Ethics and Epistemology of Scientific Research”, a research associate at the Center for Artificial Intelligence Research (CAIR), and an external collaborator at the Data and Computational Ethics research group at Stellenbosch University. He is interested in questions of machine responsibility, artificial moral agency, and the impact of technology on (human) values. Before pursuing his PhD he received a BA and MA in philosophy from Stellenbosch University.
Location: Zoom (contact Dr Andrea Palk apalk@sun.ac.za for details)
Wednesday 21 April 13:00-14:00
Title: The indispensability of collective responsibility for meaningful control of autonomous weapon systems
Abstract:
The introduction of Autonomous Weapon Systems (AWS) onto contemporary battlefields raises concerns that they will bring with them the possibility of a responsibility gap, leaving insecurity about how to attribute responsibility in scenarios involving these systems. A prominent attempt to avoid this outcome has been to insist that any such systems deployed must fall under Meaningful Human Control, which will ensure that responsibility can be adequately distributed in the case of harm. However, there are shortcomings in the current conception of Meaningful Human Control as developed in the discourse. Thus, my first aim is to provide a more adequate account of Meaningful Human Control that I entitle Sufficient Moral Control. Thereafter I will demonstrate that the richly relational context in which these systems will most likely be developed and deployed will require, if responsibility gaps are to be avoided, that Sufficient Moral Control must allow for moving beyond exclusively individualist understandings of the autonomy, control, and responsibility at work in these cases. By opening this path, it allows collectives to have Sufficient Moral Control over an AWS that forms a part of said collective, making the collective a legitimate target for collective responsibility.
Speaker:
Dr Niël Conradie is a postdoctoral researcher working in the Applied Ethics Group of the Department of Society, Technology, and Human Factors at RWTH Aachen University, in Germany. His current research is focused on collective responsibility and how this relates to AI and other emergent technologies. He earned his PhD in philosophy, focussed on the intersection of responsibility and action theory, at the University of St Andrews, Scotland. Before this he received his MA in philosophy and BA at the University of Stellenbosch, South Africa.
Location: Zoom (Contact Dr Andrea Palk for Zoom details apalk@sun.ac.za)