The danger isn’t that machines will become as intelligent as humans. That in itself is not a threat postulate. The risk is we were wrong in our anthropocentric assessment – machines didn’t need to be intelligent at all. Mirrors do not need to think to make us hate what we see.