It isn't so much that, as given we don't know and we have equal evidence in a number of different directions, our utility function should be biased towards believing in things that empower us and help us lead a better life. If the weight of evidence were in favor of a hypothesis that was disempowering, the situation might be different, but it isn't.