It’s a dream that has featured in countless works of science fiction – the humanoid robot so realistically rendered and programmed that it is impossible to distinguish from an actual human.
So far, such accuracy has eluded scientists and engineers, but recent developments by a team at the University of Pisa represent a step in the right direction. Led by computer scientist Nicole Lazzeri, a doctoral student at the institution, the team of scientists have created a lifelike “female” robot whose face contains 32 small motors that mimic the movements of the 52 muscles found in the human face, allowing the robot to simulate a range of human emotions with a surprising degree of accuracy.
The robot – appropriately named FACE – is powered by software called Hybrid Engine for Facial Expressions Synthesis (HEFES). This unique software – which has actually been around for about 30 years – is designed to mimic an entire sliding scale of human emotions. We’re not talking simple smiley and sad faces, either – FACE can express amazement, disgust, fear and surprise, and even vary the intensity or blend more than one of these to create a realistically subtle effect. The result is impressive – but rather creepy. Have a look at this video of FACE in action and see what you think:
Interestingly, that weirded-out feeling many of us get from seeing a robot like FACE, or even lifelike dolls or mannequins, has been graphed by scientists and is something the team behind FACE had hoped to avoid. Known as the “uncanny valley principle”, the feeling occurs when we see something that’s very similar to a human, but isn’t quite realistic enough to pass as human. The discrepancy makes us uneasy, and explains why people find it harder to deal with humanoid robots than ones that are obviously non-human.
The effect can theoretically be mitigated if the robot is sufficiently lifelike that our brains can’t tell the difference, and it was this that the FACE team were hoping to achieve. While I admit they’ve made a good effort, FACE still looks like something that would come in handier at a Halloween party than, for instance, in any kind of customer service role – have a look at the next video for a sense of how the latter could work.
(Incidentally, the FACE robot is modelled on the wife of one of the team members – I wonder how she reacted to its looks?)
Why does it matter if humanoid robots freak us out? The clue may lie in Dr Lazzeri’s research specialties. According to the FACE team’s website, Dr Lazzeri studies human-robot interactions, with an eventual view to developing and/or improving humanoid robots that can help people with social disabilities. Presumably the robot can’t make people uncomfortable if it’s going to help them.
What do you think of FACE? Are you creeped out or fascinated? Or a little of each?