I’m tired of having to hear and talk about AI, so I adopted a simple position on the subject which I could reflexively toss at anyone with a different take on the matter before getting on with the rest of my day. It’s a solution which I think many other theists have landed on, and, roughly, it runs like this: Intelligences are God-given and given by God alone, so at the end of the day, we don’t have to worry about whether the AIs we’ve created are conscious — no matter how much they do or don’t seem to be, and whether or not the scientists say they are — because we can rest assured that we are uniquely ensouled and that no Butlerian Jihad, no cosmic cataclysm, no day of wrath awaits us. No matter how stupid ChatGPT makes students and no matter how obsolete Claude makes employees, no matter how much AI degrades the quality of human life, it can never change the fact that I have a nature dignified by Jesus’ incarnation, passion, resurrection and ascension, whereas an LLM does not.
I’ve realized, however, that this belief is not what I actually believe — it’s a coping mechanism. If I really believed it, I’d feel some sense of comfort about whatever AI is going to do, except I don’t. In fact, I’m terrified. Why? Maybe, subconsciously, I have a hunch that AI might just make itself a soul.
Really, this belief isn’t so beyond the pale. Kids, even if they haven’t seen “Toy Story,” tend to believe their dolls and stuffed animals and action figures are ensouled and treat them as such.
More seriously, the Jewish tradition — depending on how literally you interpret it — sometimes alludes to humans creating souls. In the Talmud, Rava makes a man but apparently fails to make him a soul, although it’s implied that a sinless person would be able to. (The trouble, of course, is finding a sinless person.) Later on, Rav Ḥanina and Rav Oshaya at least manage to make a calf which seemingly has an animal soul.
There are more conventional reasons to entertain the possibility of the ensoulment of AI, though. One is alluded to above: Lots of people have a lot of fear about AI, and dismissing this collective terror by dismissing the possibility of man turning computers into fellow rational animals is dishonest to an intuition held by so many.
Another reason relates to this fact, and is a lot like what the Mennonite theologian Gordon D. Kaufman argued about the apocalypse in “Theology for a Nuclear Age” (which I bought at a Half Price Books in high school because of how sick that title is). Kaufman begged people to consider “our new-found ability to obliterate, and thus the possibility that we will obliterate, all future human life,” and to realize that this situation was unprecedented. Controversially, he then rejected the proposition “that God, as the redeemer and savior of humankind and the human enterprise, has so bound Godself to humanity and the human enterprise ... that this utterly calamitous self-destruction of humanity will never be allowed to occur.”
I agree with Kaufman that such a position obscures an evident truth: that an apocalypse which is “an act of human doing rather than of divine will” and for which human beings “are absolutely and fully responsible” is a theological and political possibility. We can’t just point to God’s sovereignty and providence before shrugging our shoulders.
For Kaufman, the catastrophe was to be one of ICBMs — now we have an AI catastrophe to worry about, too. If we smugly say, whether because we’re Catholic or because we’re Mennonite, that we needn’t worry about an ensouled AI because it simply couldn’t happen, we increase the chance that it will.
