The more human something appears, the more we are attracted to it – until a point, that is. That point is known as the Uncanny Valley, and has been causing trouble for thinkers and dreamers alike before since before it had a name.
The concept was “discovered” by Japanese roboticist Masahiro Mori. It operates on the principle that if you were to graph emotional responses to nonhuman entities, there would be a downward spike in the otherwise upward slope just before an object reaches a completely human appearance.
Although the idea was originally meant to be applied to robotics, all nonhuman items can be included in the scale, from inanimate objects to stuffed animals to animated characters. Each elicits a different but predictable level of empathic response.
To complicate matters further, there would be more variance in a graph of only moving objects – animals and robots – than in a graph of only inanimate ones.
For example, we have a low empathic response to a slug or other invertebrate – something that would inhabit the left side of the graph – because it shows almost no human characteristics. We can relate to chimps – something that would inhabit the right side of the graph – however, because they show numerous, very human characteristics and actions.
Somewhere between these two points lies an area in which a person will feel repulsed by the appearance of an object. At this stage, the item is almost human enough to be accepted as real, but still posesses recognizable defects. In a non-realistic item, humanlike characteristics will stand out and render the item appealing. In a almost-realistic item, alien characteristics will stand out and render the item revolting.
This might be attributed to the fact that in humans, abnormal functions or appearances are associated with mental illness or corpses. It is also suggested that evolution would have selected for individuals who were able to detect micro-anomalies in fellow homo sapiens, picking out defects that might pose a threat to the gene pool. We are able to detect these anomalies in figures that fall into the Uncanny Valley and are subconsciously turned off to them.
Because we live in an age of media and scientific advances, we are approaching the edge of the Uncanny Valley and attempting to cross it. Invariably, some will fall in.
The race to render the most detailed computer-generated characters has left an exceptionally high number of fatalities in its wake. At the box office, movies featuring photorealistic humans – Spirits Within and The Polar Express, for example – fare poorly. This is because of conflicting messages sent to the brain: we register the characters as humans, but are aware of the eeriness of their not quite right movements.
As technology advances, however, we are making greater leaps. Keyframing and action capture techniques make characters like Gollum from The Lord of the Rings series not only believable, but capable of evoking a complete range of emotions in viewers, from hatred to pity. A number of movies that explore the subject, such as I, Robot, Bladerunner and A.I., make ironic but successful use of computer generated imagery, themselves.
Some use it to their advantage, to evoke a sense of creepiness or unease, as in the film A Scanner Darkly, which utilizes an animation technique known as rotoscoping, or tracing over live action footage. Others avoid it entirely, as in Honda’s Asimo, a robot that is extremely lifelike in movement but whose appearance is simplified and possesses no facial features. And still others continue to find ways to leap the gap entirely.
