Log in   •   Sign up   •   Subscribe  feed icon

How Do You Make A Robot More Human? By Making It Neurotic, Apparently

Quick question, ladies and gentlemen. What is one of our defining traits as human beings? What is it that separates us from the animals and machines that surround us on a day-to-day basis. Is it empathy? Compassion? Self-awareness? Abstract thinking and the ability to plan ahead?

Apparently, it's none of the above - it's our neuroticism. Our little mental quirks, flaws, and obsessions evidently make us more real; more human than any other trait. That's a bit of a strange answer, isn't it? I thought so too.

Researchers at the University of California didn't.

These scientists have touched on neuroticism as a key factor in developing a robot that's capable of mimicking human behavior, allowing it to act more convincingly 'human' than its peers.  According to lead researcher Jeff Krichmar, they're trying to make the robot brain work more like a human brain. The key to achieving this, he explains, lies in the development of "neurobiological robots."

What that effectively means is that the researchers are looking for unique human or animal traits that can be copied, digitized, and replicated in order to make robots as a whole function more effectively. A large part of these experiments involve robots with phobias, obsessions, and 'personality quirks' - one 'bot might be afraid of open spaces, for example; another may show signs of obsessive-compulsive disorder.

To get themselves started, the team took a mouse in a cage and monitored it as they pumped varying levels of dopamine and serotonin into its brain - two chemicals that control the body's pleasure centers and well-being. After collecting data from the mouse's behavior, they programmed it into the software of the robot rodent - effectively, they replicated the mouse wigging out in software. 

"We are doing mathematical models of brains or cognitive systems, then putting it in software; it becomes the controller for the robot," said Krichmar. 

So, wait...is something like that actually going to bear fruit?

Evidently so. 

I've seen a lot of science fiction literature that depicts robots as cold, unfeeling beings with only conquest (or in some cases, liberation) in mind - automatons whose minds and objectives are so alien; so different from our own that we can't possibly hope to fathom them. Experiments like Krichmar's are something of a reality check - they serve to remind us that robots may never reach the point of self-awareness, and even in the event that they do, they're going to be more like us than we realize.

I'm not certain whether that's a good thing or a bad thing yet. I suppose we'll have to see the end results of the experiments being carried out by Krichmar's team, and all like them. Who knows what we'll end up with then? Maybe what's waiting for us at the end of this particular tunnel is a race of needy, obsessive, and neurotic robots that need us more than we need them? 

Maybe not. 

The University of California's findings were recently shared in Hong Kong at the "IEEE International Conference on Robotics and Automation."