It’s not quite Terminator, but it could be the precursor: A robot has for the first time passed a logic riddle that presumably tests for self-awareness, a test that only humans have been known to pass previously, reports Business Insider.
For the experiment, which was led by Professor Selmer Bringsjord of New York’s Rensselaer Polytechnic Institute, three robots were given a modified version of a traditional logic test often referred to as the “wise men” riddle.
In the original riddle, three wise men are vying for a chance to be the king’s new advisor. To test them, the king presents them with his riddle. He places a hat on each of their heads, such that each wise man can see all of the other hats, but none of them can see their own. The parameters of the puzzle are as follows: (1) Each of the hats is either white or blue, (2) none of the wise men are allowed to communicate to one another, and (3) it is a fair contest. The winner of the test is the first wise man to stand up and correctly declare (without just guessing) the color of his own hat.
The solution to the riddle is that all three hats must be the same color for it to be a fair contest, so whatever the color is of the other two visible hats, that too must be the color of one’s own.
For the robot version of the test, the three robots are told that two of them have been given a “dumbing pill,” a program that renders them mute. The robots were then asked if they knew which one of them was spared the dumbing pill. One of the robots, the one that was not muted, then stood up and simply responded “I don’t know.” Immediately upon uttering this phrase, the robot then realized that it did know the answer to the question, and that it was able to prove it. Since it was able to speak, it knew that it was the one spared from the dumbing pill.
Though the logic of this riddle is rather uncomplicated, researchers believe that realizing the solution requires something that robots aren’t supposed to have: self-awareness. In order to come to the correct solution, the robot had to understand the rules of the game, understand that it was a unique entity separate from the other contestants, and, perhaps most importantly, it had to recognize the sound of its own voice.
The true extent of the robot’s self-awareness, based on this test alone, is perhaps a matter for philosophical debate, but researchers on the project insist that the robot’s behavior is at least a “mathematically verifiable awareness of the self.” Amazing
You can see a video of the robot passing the test below, and see for yourself if it demonstrates proof that robots are capable of self-awareness: