The more a robot looks like a human, the more we respond to it. Because actroids looks so much like a human being we can expect it to elicits the sort of human-like responses. Osaka Unversity in Japan teamed with corporate sponsors at Kokoro and have developed ReplieeQ1-expo whose skin is made of silicon, rather than hard plastic, so it’s much more flexible. Her hands, eyes, and lips are incredibly realistic and detailed. She’s actually modeled after a young popular Japanese news anchor. (Popular because she’s so sexy?) Nonetheless, she’s very robust and can take strong physical hits, but if someone were to hit her, say in public, onlookers would most likely be upset. This sort of thing doesn’t happen to a robot that looks more mechanical, like Adam Smith’s Theory of Moral Sentiments says. But there’s a unique balance between behavior and appearance.
Behavior is somewhat difficult to model, because androids in general tend to be somewhat stuffy by default. C-3PO for example is always uptight and stuck-up. Humans tend to make more natural movements, and even unconscious actions, like blinking, adjusting position, picking our noses or scratching our heads. If androids are capable of making these sorts of movements it would feel more natural, and give actroids a human presence, with our small imperfections and annoyances. To get actroids to mimic human behavior, compressed air in her stomach powers a series of actuators in her arms and face. She can make a very natural and large movement because of the actuator’s power, but unfortunately much of the machinery to run this it outside of her petite body frame.
The picture on the left is the most recent actroid who is modelled after one of his male developers. Important gender research is taking place with the two different models. Sociological research has already found that people are less likely to lie to an actroid, or at least when they do, their eyes become smarmy and shifty as if they’re speaking to a real human being. It would be interesting to see the results of gender android studies. Or to hear how Haraway would respond to such ideas. Humorously, it has also been noted that there’s a point at which, because the actroid is so human-like, that there is a tendency at times to look zombie-like, like an animated corpse. When the actroid no longer is interesting, but just plain creepy looking, this is called the “uncanny valley”. There are a number of ways of solving this. One is to refine the technology to make her movements less jerky.
In the US, researchers at MIT are first paid for with military educational funding. Their paychecks come from the United States military. If something like this is being developed in the US it is most likely and primarily a military operation before it is a civilian operation. We’re probably secretly working on android projects for military purposes, or at least “human intelligence” usage. I would find it very difficult to believe if there is no military or governmental funding involved with actroids and android projects. The Japanese have no such obligation to their military, constitutionally. Their researchers are instead learning more about how the human brain responds to new technology by creating actroids. One of the goals of actroid development is to make technology so life-like that they can use it to understand human behavior. The human mind uses some of the most interesting and complex technology. Computer scientists are learning more about how the human mind works, functionally and behaviorally, by researching ourselves through this dim mirror which is becoming every day clearer.