There’s more public focus on how we handle hypothetical machines that are actually sentient than on how we handle the psychological problems stemming from believing non-sentient algorithms are consciously intelligent.