Fabio Tollon
Bielefeld University
In this paper I argue for a re-framing of the debate surrounding artificial agency. I do so by running through some standard problems in action theory and applying this to the case of artificial systems. What we learn from this story is that agency ought to be understood in a more practical and normative way than is commonly suggested. Agency is not something ‘out there’ in the world, waiting to be discovered. Rather, agency is ascriptive, in the sense that what we are interested in concerning agency is whether the agent is responsible for their actions. There are three major upsides to the argument I present.
In the first case, it places significance on the practical questions concerning technology. Instead of attempting to provide robust metaphysical criteria, we should instead focus on how these systems ought to be integrated into our lives.
In the second case, I believe that placing questions of responsibility at the centre of our discussions of agency allows us to draw on recent literature on ‘responsibility gaps’. Much of the literature on this topic has been concerned with insights from ethical and moral theory, and so this papers contribution is to add an action theoretical perspective to the debate.
Third, then, this discussion echoes calls for socio-technical perspectives with regards to AI-systems. Such an approach understands technological systems as not only being dependant on their ‘hardware’ but also on the human institutions and practices in which they are embedded.
Instead of asking whether artificial systems being agents is a matter concerned with the way that the world is, we can instead see the question as being about the reasons we might have for being interested in them being agents in the first place.
Chair: Sharon Casu
Time: September 6th, 14:40-15:10
Location: SR 1.003
SOPhiA 2026 