University of Tokyo, Graduate School of Interdisciplinary Information Studies, ITASIA
抄録
Smart-technologies substituting us and acting for us have become increasingly ubiquitous over the last decade. We are quickly getting used to delegating everyday tasks to a multitude of artefacts. Artefacts that constantly mediate our experience of reality and help us in the process of making decisions. However, from a moral point of view, we do not yet know how to consider all these entities.
In the social sciences, Actor-network Theory (ANT) has provided a consistent framework for the analysis of non-human agency. This has been theoretically possible thanks to the detachment of the notion of agency from that of human intentionality. However, it is not clear if and how a notion of agency detached from intentionality could also be embraced by the field of ethics. What is the usefulness of ascribing moral agency to non-human entities? Would such a new notion of moral agency be a characteristic of one specific category of entities, or could it be ascribed to anything? Furthermore, how could the question of responsibility be reframed in order to fit a nonanthropocentric ethical approach?
The paper focuses on the crash of a self-driving car, an example which is used to review advantages and limits of an ethical framework informed by ANT. Moreover, the article illustrates the alternative non-anthropocentric approach of Information Ethics (IE), highlighting the potentials of its narrower definition of moral agency. Ultimately, the paper shows that, despite their different ontological foundations, ANT and IE reach comparable conclusions in the moral analysis of the car crash: both these theories leave in fact the door open for the assessment of the moral agency of a self-driving system. This is possible due to a conceptual shift from the idea of moral responsibility to that of moral accountability, terms which, however, still lack a fully consistent definition.