Dr. Frank McAndrew

They saw a story just last week where a plane crashed, or they saw a story just last week where somebody got killed by a

And so it's right there.

If you ask somebody, how likely is it you're going to slip in the shower and fall and kill yourself?

They look at you and say, no, that's never going to happen.

But in fact, that happens every day somewhere, happens all the time.

But it doesn't make national news, right?

New Jersey man falls in shower and you don't hear that.

So they underestimate the probability of that because they can't think of an example of it happening.

I think there are different kinds of creepiness.

And some other researchers have kind of pointed this out.

And mine was all about the ambiguity of threat.

You know, is there a danger here?

But it's clear that a doll is not going to kill you.

There's another way of thinking about creepiness.

I think it's called categorical ambiguity.

The problem with creepy dolls and robots and lifelike AI images that are a little off is our caveman brains aren't prepared for things.

to look so human, but aren't human.

And so we respond to them as if they're real people, but consciously we know they're not.

And so there's this battle going on between our conscious processing of what's going on here, but our unconscious involuntary response to that.

As something looks more and more human,