The Japanese are definitely ahead of the curve in terms of developing robots – especially “humaniform” robots. But I think there latest attempt is a swing and a miss.
Meet the Telenoid R1 (yes – the next one is sure to be called R2). The idea is to create a minimalist robot for use at the receiving end of an audio communication (a phone call). The purpose of the robot is to mimic the gestures and physical expressions of the talker, so that the receiver will be better able to judge their mood and intent.
The underlying principle is sound – humans makes use of a massive amount of non-verbal communication. We talk with our hands, eyes, facial expressions, and body movements in addition to our voices. This does make it more difficult to communicate subtly over the phone.
This phenomenon is further exacerbated with type-chatting and is why the use of emoticons is common – to infuse tone into writing when it is not otherwise obvious. Sarcasm seems particularly difficult to convey.
But while the underlying concept is sound, the execution is curious and seems to be a complete failure. First, I have to ask why having a bulky (even if minimalist) robot at the receiving end is thought to be a superior solution to just having video. If you really need the visual cues, video chat on Skype or on your iPhone 4. Certainly the R1 cannot be cheap, and video solutions are likely to be more accessible.
But further, the robot is just creepy (and is widely being reported as so in the Western media – just Google “creepy robot”). Watch the video:
I don’t know about you, but I find the twitching and gyrating of the robot to convey no useful information, and to be distracting if anything.Perhaps it would be effective if the technology worked much better, but it seems likely that by the time such a device can be perfected to the point of usefulness, video phones will be widely available. It seems unlikely, therefore, that at any point in the future talking to an animatronic robot will be in common use.
I understand this can largely be just a proof of concept, not something actually intended for use. It is part of the technological development of human-looking robots that people feel comfortable with because they look and act like us.
Another thing to consider, however, is the uncanny valley – the hypothesized range of human simulation that is too realistic, but just not realistic enough. Below a certain level of realism in depicting a human, the figure will seem cartoony and fake, but in a comfortable and pleasing way. Increase the level of realism to a certain point and human depictions seem too real, while not be quite realistic enough. This seems to trigger an emotional reaction of disgust or revulsion. Such depictions are creepy. Perhaps they are real enough to seem like abnormal humans, rather than artificial depictions of humans.
The R1 may be flirting with the uncanny valley, but also I think the minimalist approach to the face and body makes it look like Casper the robot. It’s just creepy.