Abstract

This paper offers an Aristotelian-Thomistic response to the question whether AI is capable of developing virtue. On the one hand, it could be argued that this is possible on the assumption of the minimalist (thin) definition of virtue as a stable (permanent) and reliable disposition toward an actualization of a given power in the agent (in various circumstances), which effects that agent’s growth in perfection. On the other hand, a closer inquiry into Aquinas’s understanding of both moral and intellectual virtues, and a more detailed analysis of the ontological status of AI, show that it is highly unlikely to envision the design of specifically human-like reason-based and/or behavioral-based (“strong”) AI that would possess properly human virtues. Still, virtuous “weak AI” might be possible, although a question ought to be asked whether we should classify artifacts’ virtues using categories developed in reference to specifically human dispositions and actions.