McAnulty College and Graduate School of Liberal Arts
Ethics, AI, Intelligence
Smarter people or intelligent machines are able to make more accurate inferences about their environment and other agents more efficiently than less intelligent agents. Formally: ‘Intelligence measures an agent’s ability to achieve goals in a wide range of environments.’ (Legg, 2008)
In this dissertation we extend this definition to include ethical behaviour and we will offer a mathematical formalism and a way to estimate how ethical an action is or will be, both for a human and for a computer, by calculating the expected values of random variables. Formally, we propose the following measure of ethicality, which is computable, or at least can be approximated: The ethicality of an action can be measured as the change in ability to reach a goal of all agents affected by an action, for each action taken, and weighted by the intelligence of the actor.
Thus, we claim that agents cannot be morally responsible for consequences they are not smart enough to infer. They are (morally) responsible only for what they could foresee given their intelligence. Finally, intelligence provides us with estimated and exact measures: The former is used by the actor who must pick a course of action, especially in the heat of the moment which would require a lot of estimations to be made, whereas the latter can be used by any agent with sufficient computational power and time and should deterministically yield the same results.
Napolitano Jawerbaum, A. J. (2023). Proposing a Measure of Ethicality for Humans and AI (Master's thesis, Duquesne University). Retrieved from https://dsc.duq.edu/etd/2167