Abstract
The article by Howick et al.1 is a welcome exploration of the practicalities of empathy cost-effectiveness; though, I would like to introduce an alternative narrative of empathy in robotics. The uncanny valley effect postulates that with an increasingly humanoid appearance of a robot, the human observer’s response becomes increasingly empathetic until a threshold of uncanniness is reached. Triggering vulgarity.2 Uncanniness, it must be said, is defined as ‘eerily similar’, rather than ‘mysterious’ per se. Therefore, artificial empathy is an aloof concept to modern day and even near future robotics.
Empathy is not mere communication; it’s something more. In fact, it’s simulative perspective-taking: the ability to anticipatorily ascribe and decode the behaviour of another through the activation of mirror neurons. Therefore, the aforesaid ‘something more’ is not a mystical phenomenon but a physiological process. If robots do indeed remain more machine than man, then empathy is lost and not to be attained due to a lack of physical similarity. And any attempts to convert the machine form to that of mankind: the uncanny valley effect. It appears artificial empathy is in a state of paralysed paradox. This is further solidified by the sorites paradox: the intermingling of human and non-human traits destabilises our sense of human identity. In an emotionally charged healthcare setting, greater demand is placed on the empathetic consultation, supplying this from a robot (humanoid or not) will likely fail.
So, the authors’ view that robots should be trained in empathy is an interesting position. Though, perhaps it’s the humans that need artificial empathy training after all. Oh, humanity is indeed a complex thing …
