Normative Requisites of Forgiveness and AI Agency

Regarding forgiveness, there is disagreement on the following “contrast pairs.” (i) The victim (the person who forgives) forgives the offender (the person who is forgiven) for doing something only if it is morally wrong for the offender to do this thing; or even if she has done no wrong, the victim forgives the offender only if she takes herself to have done wrong. (ii) The victim forgives the offender for doing something only if she is blameworthy for doing this thing; or even if not blameworthy, the victim forgives the offender only if she takes herself to be blameworthy for doing this thing. In this paper, first, I appeal to “Frankfurt examples” to argue that the belief options – the second members – of each contrast pair are more tenable. Second, I discuss why even these belief options threaten forgiveness if determinism is true in virtue of determinism’s imperiling wrongness. Third, I broach the issue of whether artificial intelligence agents (“AI agents”) can be apt candidates of forgiveness. I synopsize conditions AI agents must satisfy to be fitting subjects of forgiveness, emphasizing agency requisites for blameworthiness and wrongdoing. The summary’s principal aim is to animate further inquiry into whether AI agents can be agents of the sort that forgiveness presupposes.

کلیدواژه ها:

Artificial intelligence agency ، blameworthiness ، Frankfurt examples ، Forgiveness ، wrongness