I think philosophically, yes, it doesn't really tell us anything interesting because no sentient human would choose nuclear war.
However, it does work as a test case for AIs. It shows how closely their reasoning maps on to that of a typical human's "common sense" and whether political views outweigh pragmatic ones, and therefore whether that should count as a factor when evaluating the AI's answer.
However, it does work as a test case for AIs. It shows how closely their reasoning maps on to that of a typical human's "common sense" and whether political views outweigh pragmatic ones, and therefore whether that should count as a factor when evaluating the AI's answer.