On Algorithmic Opacity and Moral Certainty

Abstract

In modern society, algorithms play an important role in social and cultural realms, in political and economic spheres. In spite of algorithmic pervasiveness in many areas and wide diffusion in digital life, algorithmic opacity is still poorly understood compared to other ethical issues (e.g., fairness, accountability, and transparency). In this essay, we try to elucidate the relation between algorithmic opacity and moral certainty from the individualistic standpoint and through the virtue ethic perspective. For doing so, we follow hermeneutic tradition and rely on interpretation of recent authors and impactful papers. We summarize our argument as follows: if the algorithm is understood as the combination of rules and numbers we create for simplifying our lives and sharing with others, then our present activities and future actions as imagined, realized or missed, ascertain if algorithmic opacity become a moral issue or problem for us and others. Among the implications, we emphasize that sometime dormant and hard to anticipate, algorithmic opacity becomes an apparent during executions, deployments and prolonged uses of algorithmic systems. Moreover, our lived experience and disharmony between our unrealized expectations and unanticipated algorithmic behavior may lead to moral issues and problems for us and others. Overall, algorithmic opacity may constantly evade the formalization efforts (e.g., outlining as guidelines, principles) or quantification exercises (e.g., assigning numerical values to symbols or signs), both of which are essentially social practices.