The face of robots

National Museum of Scotland, Edinburgh, 2019

The ambition of the exhibition is contained in the museum’s statement: “From early mechanised human forms to today’s cutting-edge technology, this major exhibition reveals our 500-year quest to make machines human.” There is more to this than it appears: why on earth should robots be human-like, if their purpose were to mechanise certain tasks, or even to replace humans for the performance of certain repetitive or dangerous tasks, or indeed tasks which humans perform too expensively, or without enough precision or consistency ? And any visit to a car factory will confirm that most industrial robots do not look human, except by metonymy as one speaks of their “arms”, for instance.   Why should a robot which helps extinguish fire look like a fireman ? Or a killing machine like a soldier ?

But we have tried indeed for a long time to replicate humans, even when this replication was useless from any practical purpose: the early human-like robots are just machines which vaguely move articulated arms and legs, without being able to accomplish any meaningful task. Science fiction has produced an endless stream of images of such machines, and the innumerable “soldiers” of the many “dark forces” of literature and cinema, but also of video games, are human-like, to some degree. Ugly versions to be terrifying, or very similar to be worrying, until the ultimate evolution which merges robots with humans in a single artefact through the myth-to-become-reality of bioengineering, of which the 1982 Blade Runner is perhaps the most convincing precursor. Let us not forget that the action is set in the Los Angeles of 2019…

Now that technology has moved forward, we are capable of making useful human-like robots, which can actually walk or even jump, despite the immense complexity of achieving this movement through electro-mechanical means, engage into simple conversations thanks to AI algorithms and good voice replication, and even look attractive thank to the advanced replication of skin texture and 3D modelling.

The question I would like to briefly address here is that of the face of robots. From the perspective of their maker, the urge to act as a god is pretty obvious, and as old as the most ancient mythology. The clearest image is that of the Golem. As the God of Genesis created man in his own image, man tir

From the perspective of their users, or human “companions”, the only reason why robots should have a face is to be able to interact with humans as if they were themselves human. And the reason why we would like them to be human-like in such an interaction is unlikely to be only for the superficial comfort that we feel in their company, but because we are perhaps inclined to replace potentially troublesome humans with ones which are less problematic. The ideal servant is one that does not eat, sleep or complain; the same went for the ideal worker, except that nowadays a worker without initiative and proactive contribution to the common venture has become a useless worker. Robots with a human face can also be endearing, in particular for children, and thus contribute to the recreational function which is becoming strategic in the realm of social control; they also seem less threatening, under certain conditions, which opens the immense market of basic care, whether of children, the ill, or elderly people.

Mori curve

This brings us to the famous theories of the Japanese robotics scientist Masahiro Mori, which date back to 1970, about humanoid artefacts; according to Mori, humans feel a growing sense of familiarity up to a certain point of resemblance of the robot with humans, say around 75%, after which exists what he calls the “uncanny valley”, where the appearance of an artefact which looks nearly human creates a deep sense of discomfort or even fear. Then, when getting closer to actual human traits, familiarity is again high, and growing; it goes through steps which include the puppet, the ill or ugly person, and ultimately the healthy person; beyond a point which we could name that of an “average” healthy person, Mori hypothesises a kind of symmetrical curve. Some faces of the Buddha, let us say, are beyond the human, and do generate a feeling of estrangement, or at least a distance. But it is also a common experience that extraordinary human beauty, when met face to face, arouses a degree of uneasiness. This is why robots are always kept one side or the other of the “uncanny valley”, mostly on the left-hand side such as with Ribina, Romeo, Pepper, Asimo or iCub, but also on the right-hand side with Actroid-SIT or Actroid-DER2, or Harmony. The robot Ibuki which is “on the edge of the cliff” is indeed slightly worrying…

There is merit in this theory, as it matches actual human response to humanoids. However, it does not really explain the trend in multiplying humanoids, which again are not indispensable as such.

The key point I would like to make is that we reached a time in history when, to put it in the extreme, man seems to want to get rid of man. It is easier to interact with a machine, easier to control it, and the psychological remuneration provided by a realistic enough proxy could become satisfactory for generations which are already raised in the partial company of virtual or image-mediated worlds.


It is fascinating to see how the “face” of robots has evolved from a sort of metal helmet with a nose and flashing eyes to a white or grey plastic globe, or quasi-globe, during the first decade of this century, and now seems to venture in the territory of “imitation-flesh” with features very close to those of actual humans, particularly in the case of sexual robots, for obvious reasons, as intimacy requires maximum familiarity, and a focus on all the stimuli of attraction.

Can one forget that the machine is a machine, and if not, why would we seek any form of emotional interaction with a humanoid? Cinema has attempted to provide an answer. It all started with Metropolis (1927), went on to Blade Runner (1982), then the remarkable Swedish TV series Äkta människor (Real Humans, 2012), all the way to Ex Machina (2015).  Clearly, resemblance to humans, and more importantly the view of human beauty, to the extent that it is coupled with the AI required for likelihood, is an effective instrument of distraction, a distraction which opens the door to emotions.

We can therefore sight a partition between “assistant-neutral” robots, with a face which is pleasant but not human, quite to the left of the uncanny valley, “emotions-triggering” robots with a human face, designed to generate some form of emotion, well to the right of the valley, and “performance-driven” robots capable of amazing mechanical and recognition activities, such as Atlas or Handle of Boston Dynamics, which are clearly un-human as they have no faces, but mimic many human attitudes such as jumping or walking down the stairs, and bring us to an animated version of the faceless universe of De Chirico; this universe is one of amazement and strangeness, and could indeed be the closest to the uncanny valley. The idea of robots with hypnotizing capabilities is already in the air…

Number oof robots in operation from 2000 to 2017 worldwide

The faces – or absence thereof – of robots are part of a contemporary iconography, and can be related to different worlds of meanings, of purposes, of social relationships. There is a potential for ever more creativity in the appearance of robots, and perhaps in an aesthetics of attitudes half way between choreography and the imitation of natural movement. One cannot avoid to relate form and substance when evoking robotics: at what point does form start conjuring up a deeper substance, an inhabitant to such form? As the number of robots progresses exponentially, we are to come more and more often face-to-face.

Laisser un commentaireAnnuler la réponse.

Quitter la version mobile