The uncanny baby valley

I can’t wait for the day when robots and humans will co-exist…or when robots will take over the planet, Matrix-style.

Until that time actually occurs, I’m pretty sure we have nothing to worry about because people in general are still scared. There’s this concept in psychology and robotics (I think those are the appropriate fields…) called the Uncanny Valley. For those unfamiliar and unwilling to click on the link, it’s the concept that as robots/figures/other objects approach the look of a healthy human, the more we feel familiar with it – comfortable, not afraid, not worried about its actions.

However, there comes a point when the other appears TOO human. At this point, our comfort with the other drops…by a lot. It is this drop that is the uncanny valley. Example – a zombie or a corpse appear more human than a robot, but we feel much less comfortable with these two because they are too close to humans, yet very clearly aren’t. It’s interesting to think about how humans feel comfort in the familiar, yet when things are too familiar we recoil.

This can be seen in peoples reactions to things beyond objects. When someone we barely knows acknowledges us at a bar, we don’t really make much about it. However, when they are overly familiar with us – addressing us by name, asking personal questions they shouldn’t be asking, we feel uncomfortable. Odd, isn’t it? You’d think we would feel better that someone is trying to be approachable.

Okay, long detour…lets get back to the robots! Is it well accepted that babies learn best by being socialized? I feel like I read that in some study before….if anything, this study proves it. It appears that some researchers have decided to start training babies to be familiar with their future robot overlords. By getting the child comfortable in their environment, then introducing a robot figure as if it was human, the researchers have been able to convince babies that robots have some sense of intelligence or thought.

If, once again, you are too lazy to click through for the article, here is a brief summary: researchers took babies, half of whom were presented a robot with which the researcher interacted with to make it appear to have sentience, while the other half were presented a robot with whom no one interacted with. The babies were witness of this interaction.

The babies who saw the researcher interact with the robot were more likely to act as if the robot itself was a real person and pay attention to its behavior. For example, if the robot were to look in one direction, the baby was more likely to look in that direction if it saw the interaction.

To me, the most interesting part of the article is that this shows our social judgments/skills can be basically trained, even at a young age. That, and we just need to get babies used to robots, but also to be cautious so that robots don’t get to smart and take us over.

Source: Human or Robot? Ask the baby

Leave a comment

Your email address will not be published.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.