Blind Justice Emotions, Empathy, Algorithms, and High-Stakes Decision-Making in Policy Execution

Abstract

Reaching the correct conclusion in high-stakes decisions in policy execution is morally important. Making the wrong decision will be either unjust for an individual citizen or for society at large. Algorithmic decision aids can potentially help us achieve more accurate and objective outcomes. However, algorithms are known for their issues with bias and opacity. Human decision-making can also be seen as opaque and prone to bias. This raises the question of which is preferable for high-stakes decisions in policy execution. Algorithmic decision aids or human decision-makers? In this thesis, I will argue that, contrary to common beliefs, emotions and empathy are essential for high-stakes decisions in policy execution. Among other things, emotions and empathy are necessary to gain the contextual understanding on which to base these important decisions. Algorithms are not capable of experiencing genuine emotions and therefore lack the ability to empathize. Consequently, the use of algorithmic decision aids in high-stakes policy decisions should be strictly limited. This limitation means that algorithmic decision aids should not be used to carry out tasks with real-world consequences. However, their use should not be dismissed altogether, as they can provide factual information on which high-stakes decisions also depend, and their utilization can be of great benefit as long as the essential contributions of emotions and empathy to high-stakes decisions in policy execution can be secured.