The Virtual Ensoulment Hypothesis: Can Virtual Individuals Be Ensouled?

Brian Cutter (2025) claims that AIs can be ensouled under right conditions which he calls the AI ensoulment hypothesis. In this paper, I extend Cutter's argument and suggest that virtual individuals can receive a soul under the right conditions, which I call the virtual ensoulment hypothesis. First, I raise the argument from analogy according to which virtual individuals can receive a soul given that they are functionally equivalent with real individuals: they have access to their environment, they have a complex behavioral repertoire and they can make inferences which are partially drawn from their memory. However, I raise the counter-argument from value which doesn’t arise for the AI ensoulment. According to the argument from value, virtual worlds do not possess certain values which make virtual individuals fit for ensoulment. I consider three ways in which virtual worlds lack value and show that two out of three of them fail. However, the third value–the value of structural stability–can create trouble for virtual ensoulment because virtual worlds lack such value. By structural stability, I mean that virtual worlds are unstable given that they depend on the computer that runs them which can have negative effects on the lives of virtual individuals. I ultimately conclude that we should be agnostic about virtual ensoulment due to structural instabilities, though I suggest that our credences should mildly favor virtual ensoulment.