Michael DeVito

Questions concerning the moral standing of artificial general intelligence (AGI) have recently drawn much attention from scientists, philosophers, and the general public. One central issue is the difficulty (perhaps impossibility) of determining whether, if AGI existed, AGI would be ensouled. In this paper, we take a decision theoretic approach to the issue. Specifically, we argue that if AGI were to ever exist, even if we thought the probability of AGI being ensouled was low, we should take the course of action as to treat AGI as if it were ensouled.

The reductionist and anti-reductionist debate features prominently within the larger canon of literature devoted to the epistemology of testimony. Reductionism is the thesis, roughly, that testimonial knowledge is reducible to some other, more general, form(s) of knowledge. We take anti-reductionism’s thesis to be the denial of reductionism. In this note, we develop an attempted counterexample to reductionism by way of discussing animal testimony. We don’t mean to suggest that the reductionist is without a response to the counterexample proposed, but we think it’s about time epistemologists start focusing on how broad we should understand testimony and to what degree, animals might guide our process in developing our epistemology of testimony.