Did This Little NAO Robot Just Become Self-Aware?

Artificial Intelligence  

Researchers have run a test of artificial intelligence self-awareness by putting a trio of NAO robots through the classic 'wise men puzzle' test of self-awareness - and one of them passed.

Scientists from the Rensselaer AI and Reasoning Lab in New York have now announced what might be a significant breakthrough toward developing artificial general intelligence (AGI). In an experiment conducted on a group of NAO robots, indicates, at least at a rudimentary level, that the robot possesses a form of self-awareness.

The findings are likely to trigger debate over what the experiment really proved.

In the study, conducted by Selmer Bringsjord, three NAO robots were subjected to an updated version of the classic puzzle known as “the king’s wise men.”

Related articles Robot Taught To Recognize Own Arm Movements In A Mirror  The Awareness  Singularity's Reality Distortion Field

The original version of the puzzle goes like this: A king calls the three wisest men in the country and puts either a white or a black hat on their heads. They can all see each other’s hats, but not their own, and they’re not allowed to talk to each other. At least one of them is wearing a blue hat. Whoever is smart enough to work out the color of the hat they’re wearing becomes the king’s new adviser.

The classic riddle presents three wise advisers to a king, wearing hats, each unseen to the wearer. The king informs his men of three facts: the contest is fair, their hats are either blue or white, and the first one to deduce the color on his head wins.

The contest would only be fair if all three men sported the same color hat. Therefore, the winning wise man would note that the color of the hats on the other two, and then guess that his was the same color.

In the updated version of the puzzle, that can be seen in the video below, Bringsjord programmed the robots to think that two of them had been given a “dumbing pill” —one that would prevent them from speaking. To do this the researchers, muted the robots by pressing a button on top of their heads, while the third one had a placebo button.

"I know now. I was able to prove that I was not given a dumbing pill."

When the researchers asked the NAO robots which pill they had received, only one of them—the one that had not been muted—stood up and said “I don’t know.”

Then, a moment later, when its program realized that it had heard its own voice, the robot waved its hand and said: “I know now. I was able to prove that I was not given a dumbing pill.”

According to the researchers, the ability to recognize the sound of its own voice and logically conclude that it had not received a dumbing pill—because it could speak—showed that it had the ability to link this realization back to the original question and come up with the right answer, pointing toward a mathematically verifiable awareness of the self.

SOURCE  Tech Radar

By 33rd Square

Embed