In 1970, Masahiro Mori, a robotics professor at the Tokyo Institute of Technology was designing a robotic hand and described his own eerie and uncomfortable feelings in response to the robot hands realism. Mori later proposed that the more human like a robot became, the more an person’s affinity for the robot declined. Subsequent research revealed that this was not just a learned human response, but that monkey’s were similarly averse to realistic synthetic faces – they would spend more time looking at unrealistic faces and real faces, than they would realistic synthetic faces. Mori questioned why humans beings are equipped with this eerie sensation and wondered if it was a part of a self preservation instinct.
This eerie sensation when faced with a robot, cartoon or doll that approached human like realism, was later labelled “uncanny valley” in a 1978 book, Robots: Fact, Fiction, and Prediction, and could be described as the trough in likability that occurs when humans find a robot to be increasingly unlikable as it approaches realism. Positive responses to robots and animations suddenly turn to a sense of revulsion as it becomes too close to real.
A few theories around the possible causes of the “Uncanny Valley”:
- Mate selection: aversion could be caused by an evolved cognitive mechanism that recognizes and immediately causes us to discredit a potential mate who is clearly “wrong”
- Mortality salience: A realistic looking humanoid robot could remind us of our own mortality, a fear that we are soulless machines of that death is inevitable. They may also elicit a fear of being replaced.
- Pathogen avoidance: That creeped out feeling an almost realistic looking robot elicits could because of a cognitive mechanism that seeks to protect us from illness, by emphasizing the small differences that could indicate disease and causing aversion
- Violation of human norms: When a robot looks like a robot, we measure it by robot standards – when it looks more like a human, we measure it by human standards, and it falls short – causing the eerie “uncanny valley” feeling
- Religious definition of human identity: Human beings tend to want to feel or be unique and special, and robots emulating humanity threatens this concept. We are also raised on warnings from folklore and stories where human like creatures lack human empathy and spirit and ultimately lead to conflict, loss or disaster, such as golem’s in Judaism or fairytale changelings.
- Conflicting perceptual cues: our brain’s don’t know whether to classify the realistic robot as “human” or “robot”, and this creates perceptual tension and “cognitive dissonance”.
- Threat to humans’ distinctiveness and identity: robots challenge our sense of social identity – how can humanity be defined as unique when it can be replicated by machines?
Is bridging the uncanny valley the ultimate design challenge? Or is it a warning to humanity? Perhaps we are all living Battlestar Galactica…we have been been down this road before, and our folklore is trying to tell us something?
Looking through these possible theories, the human response is likely a combination of several of these – a combination of evolved cognitive mechanisms and learned cultural and religious fears. I don’t know if monkey’s have culture (a question for a primatologist on another day) but they exhibited a similar aversion, suggesting the uncanny valley sensations are a relic from a bygone era. But perhaps they can also be taken as a warning for an era to come?