For those of you who haven't heard of the uncanny valley, it is from the fields of robotics and computer animation, and proposes there is a dip in the graph of human reaction from positive to negative, as a human replica approaches but not quite equals human likeness. In layman's terms, the more something looks like a human, the more positively we react to it, right up until it looks almost but not quite human, in which case it now causes revulsion.
This graph shows that hypothesis:
Taking a look at the examples given, the graph appears to support the hypothesis. Teddy bears are cute. A zombie not so much.
But on the other hand, while zombies are much closer to human appearance, they signify many things that cause negative feelings. They represent death, decay, a glimpse at a horrifying afterlife, all things which significantly impact our feelings toward them, no matter how cool and trendy they've become recently.
So let's look at something that doesn't represent those things. Let's take a look at something actually designed to attract and cause positive feelings.
Let's meet an "actroid":
Now that's kind of freaky. We're fully into the depths of revulsion that is the uncanny valley with her. I don't know what kind of terrifying visions instantly raced through your head watching that, but if they're anything like mine, you definitely believe the validity of this hypothesis. She's fascinating, and awesome, and inspiring, and a little bit of nightmare fuel.
As the technology develops, these actroids start to gain humanity, but still exhibit signs of that uncanny valley. As we can see in this next video, filmed in 2008, spontaneous interaction with humans is far closer, but still has a ways to go:
It's not limited to robotics, either. Cleverbot is a fascinating (and highly addictive) experiment with virtual intelligence. Cleverbot interacts with you, responding with original "thought" to what you type into it. It actually learns from interaction with humans, which is easily seen by asking it questions on socially popular topics. Sometimes it seems you are interacting with a real human, but if you type long enough, the artificial intelligence will show through, sometimes sending you straight into that valley.
And that brings us to CGI, and the advances made in this field along these lines. Let's meet Emily, a virtual intelligence made with computer graphics by Image Metrics:
She's almost perfect. Very, very close to human, and in fact if you're not really looking for it you can easily mistake her for human. Right up until that uncanny valley shows up and gives you subtle hints that something just isn't right with her. You can't put your finger on it immediately, but it's there at the back of your mind.
Will we ever cross that uncanny valley with robotics, AI and CGI? I'm certain we will. The rate of technological advances is astounding, and is growing at an exponential rate. I'm sure that in a few short years we'll be there, with this sort of AI interacting with us on a regular basis. From search engines to GPS navigation systems to all kinds of interactive learning, the possibilities are endless.
For now, I think we're still in the valley, but we'll be coming out the other side before long. And that's when things will start to get really scary.
Update: I was asked by the fine folks at Curiosity Quills to do a follow-up piece to this post on the uncanny valley as it pertained to speculative fiction. You can read it here on their blog. It turned out well. And by well, I mean it gives us a very chilling look at what could be a possible future for mankind.