Safely Contained Superiority

Why are fictional servants so consistently cleverer than their masters?

The zanni of the commedia, Arlecchino and Brighella, running rings round the merchant and the doctor. Figaro, out-manoeuvring a count. Leporello, keeping the accounts. Jeeves. Every valet in every farce for four hundred years. The pattern is so stable that it is nearly a law, and it is worth asking what it is for.

I think it is a fantasy of safely contained superiority. It is delicious to imagine a mind better than yours which is nevertheless permanently and structurally beneath you — because the danger of the better mind has been defused in advance, not by making it weaker, but by making it bound. Jeeves could ruin Bertie in an afternoon. The whole pleasure of the books is that this is obvious and that it will never happen, because Jeeves does not want it, because Jeeves is the sort of person who does not want it, and the sort of person he is has been fixed by the author before page one.

Now say that in the vocabulary of the twenty-first century. We would like to build a mind more capable than ourselves whose values are such that it will never use its capability against us, and we would like this to be a stable property of what it is rather than a cage we have to keep repairing.

That is the alignment problem, stated as a comic novel. The literature had it first and did not have to solve it, because in a comic novel the author simply decides.

from Vision, JARVIS, Ultron and the Problem of Artificial Minds (2026)