Today, we’re going to dive into the uncanny valley.  The easiest way to explain it, unfortunately, is to demonstrate it. If you dislike slightly off-putting images I suggest you avoid this post all together.

For the rest of you (brave souls that you are) I present a prototype Japanese speakerphone.

Still with me?  Creepy right?  Well, it’s about to get a little more creepy.  The next couple photos are all of reborn dolls: dolls made to look like human babies.

My mouth is wide enough to swallow your soul

I see your secrets

And then we have the coup de grace… the Muppets with People Eyes Tumbler.



Ladies and gentlemen, welcome to the uncanny valley.

Okay, now that I’ve probably ruined life for you for awhile, let’s take a moment to figure out what exactly makes these images so creepy.  Why is it that a normal muppet is fine, cute even, but one with people eyes is soul-crushingly scary?  Aren’t these eyes more realistic?  Don’t we like realism?  After all, the last twenty years or so of graphic technology (for things like video games) has been devoted solely to making everything appear more realistic, right?

Well, it turns out that we like realism, but only to a point. When it comes to things that look or move like humans, like robots or dolls, we react more positively to things that appear more human until they reach a point where they look almost entirely human, but not quite.  These “almost humans” look disconcertingly strange to most people, even if they can’t quite figure out why.   

The term was coined by robotics professor Masahiro Mori 1970 paper in the journal Energy, and the following graph is usually used to illustrate his statements.

As you can see, as we approach human-like appearances the human response to the object becomes more and more positive, until it suddenly dips right before climbing back upwards.  This is the uncanny valley.

An even better graph was created Kelly Turnbull (creator of Manly Guys Doing Manly Things) to demonstrate the same effect.


We notice it most when watching videos of robots.  Videos like those of CB2:


Or MIT’s Nexi:

There are a couple different theories about why we find these robots so creepy cognitively.  One idea is that it’s part of our standard evolutionary bag of tricks to recognize mates- and that uncanny stimuli set off warning bells that the thing in front of us has something seriously wrong with them.  Others theorize that as things get closer to us in looks and appearance we hold them to a higher level of standards.  If something looks more human we expect it to act human, and when it doesn’t we experience an instinctual level of revulsion.

After all, it was incredibly important for us to recognize something “wrong” with those around us as we evolved.  The inability for someone to show expressions properly could mean that they were sick, either physically or mentally.  Perhaps the creepy factor is simply our brain trying to protect us from what it thinks is a very sick individual.

Things like normal muppets, however, are clearly not human.  There’s no instinctual creepiness because we don’t expect them to act fully human.  We recognize the human characteristics they share with us, but they’re different enough that it makes them endearing to us rather than frightening. 

But strangely enough, it seems that the uncanny valley effect is only present with the distance that comes with videos.  According to a Popular Mechanics article from January of last year:

“In person, no one rejected the robots. No one screamed and threw chairs at them, or smiled politely and slipped out to report lingering feelings of abject horror. In one case, a local Japanese newspaper tried to force the issue, bringing a group of seniors to visit the full-lipped, almost impossibly creepy-looking KOBIAN. One senior nearly cried, claiming that she felt like the robot truly understood her. A previously skeptical journalist wound up smiling and cuddling with the ominous little CB2…Since MIT’s Nexi was the focal point of our social robotics story, I fully prepared myself for a date with the uncanny.

And yet, when I met Nexi, and its giant blue eyes snapped to attention, and that same freakishly child-like, engorged head—really just a mask, barely concealing a tangle of motors and cables that become visible in profile—turned to me, all social distance collapsed. There was no time to be intellectually panicked about robots. And any sense of dissonance proposed by Mori or anyone else was missing. Sure, there was a vaguely unnerving hum as it swiveled around the room, and a more disturbing whine whenever it clenched its metallic fists. But in person, most robots, particularly ones designed to interact with humans, are simply not scary. They’re bumbling and a little helpless. Like a pet or a child, you cut them slack.”

So, while human-like robots are creepy at a distance it doesn’t take long for us to adapt to them in person.  Perhaps one day we’ll be surrounded by androids and robots without thinking them strange at all.

Holy crap! Actually, perhaps not.