The Uncanny Valley
INTRODUCTION
Shown above are pictures of two robots, designed to look similar to a human. What emotions do you feel when looking at them? Some might be apathetic. But what if you feel a creepy, ‘uncanny’ sensation? If you fall into this category, you aren’t alone.
A vast majority of people would find these robots unsettling. Me personally, they make my skin crawl. Something about their face just seems off, repulsive, maybe even ominous.
How can a picture of a humanoid robot elicit such reactions? This is actually a phenomenon that has been observed and applied to in the field of robotics and artificial intelligence. It is known as ‘the uncanny valley’.
The concept was first introduced by roboticist Masahiro Mori in the 1970s. The term uncanny valley refers to the unsettling and often repulsive feeling of encountering robots that appear almost, but not quite, human.
Take a look at the graph above. As the appearance of a robot is made more human, emotional response to the robot becomes increasingly positive and empathetic, until it reaches a point beyond which the response quickly becomes strong revulsion. This is the point where the artificial figure tries, but fails to mimic a realistic human. Then positive feelings bounce way back up when a normal human is shown.
An industrial robot, with a mechanical exterior of bolts, nuts, and steel wont necessarily freak you out, but a conversing, sentient AI robot with a botched up face of a human might.
So that’s the uncanny valley in a nutshell. But the question still remains — why is this even a thing? What causes the uncanny valley effect? There are a multitude of theories, and let me tell you, they are pretty interesting.
THEORIES
Evolutionary Adaptation — Evolution rewards people who keep themselves safe, atleast long enough to reproduce. The uncanny deviations found in these robots’ facial features might activate a cognitive mechanism that pressures us to avoid potential sources of pathogens, disease, or bad genetics by eliciting a disgust response. Such a strange looking ‘human’ might have physical markers that indicate low fertility, poor hormonal health, or ineffective immune systems.
Violation of Expectations — We have a well-developed ability to perceive and interpret human facial expressions, movements, and behaviors. When a robot falls short of replicating these characteristics convincingly, our brain detects the discrepancy, and struggles to place it in a clear category — human or non-human, leading to a type of cognitive dissonance. This leads it to a feeling of unease and discomfort.
Threat to human distinctiveness — Humans might believe that these new machines are undermining our sense of human identity, and pushing for a redefinition of humanness. This poses a threat to our ego and competitiveness as a species. The more a robot resembles a real person, the more it represents a challenge to our unique social identity as human beings.
FIXING THE UNCANNY VALLEY
A number of design principles have been proposed for avoiding the uncanny valley.
Don’t mix and match — A robot may look uncanny when human and nonhuman elements are mixed. For example, a robot with a synthetic voice is less eerie than a robot with a human voice. This is because it matches our expectations.
Embrace Stylized or Non-Human Aesthetics — Rather than striving for a hyper-realistic human appearance, consider embracing stylized or non-human aesthetics for robots. By intentionally highlighting the robotic nature of the design, you can create a distinct and appealing visual identity that reduces the discomfort associated with imperfect human imitation.
Consistency — Maintain consistency between a robot’s appearance and behavior. If a robot has a simplified or mechanical appearance, ensure its movements and behaviors align with that aesthetic. Conversely, if a robot has a more human-like appearance, try to make sure it has the most realistic movements, gestures, and facial expressions possible.