Researchers Create Childlike Artificial Intelligence In Second Life
A group of scientists from the Rensselaer Polytechnic Institute have announced that they've created a logic based artificial intelligence and modeled it in the virtual community Second Life. The intelligence takes the form of a small boy, Eddie, and he "has his own set of beliefs, and the ability to reason about his beliefs to draw conclusions in a manner that matches human children his age." Plus, he's probably super creepy.
Although for now he's busy freaking out furries, the researchers have big plans for the technology, describing a Star Trek-inspired "holodeck" scenario:
[I]magine a hostage situation: How do you prepare for negotiating with a terrorist holding a hostage? Now, it's textbook and playacting. But what if you could enter the holodeck and match wits with a synthetic character that has the ability to reason in earnest about your mind, and about what you're trying to do?
Of course, just be sure that you don't make the AI too smart, or else it just will take over the entire ship and demand you find a way to turn him into a real person.
After the break, watch a video of Eddie taking part in a false belief test, proving that he's just as dumb as any four year old. Science is amazing!
Bringing Second Life To Life: Researchers Create Character With Reasoning Abilities of a Child [Rensselaer Polytechnic Institute]
[via: Destructoid]
Dawdle
Posted to Miscellany on 15 March 2008
· Comments (7)
Other stuff like this: ai, holodeck, lindon labs, second life, weird science
Just keep a game of tic tac toe handy in case you need to make it play with itself.
The problem with all of this taking place in second life though, is that the AI can't possibly deduce from what happened that there is a requirement to be present to know where the bear is. He actually answered the question 100% correctly, since Micah knows that the test consisted of the bear being transferred between boxes. Effectively then, if he actually were to look for it, he'd know that it is in box B. The AI is reading his mind :D
Since second life is about as close to reality as Uwe Boll is to being a director, none of our conclusions actually make sense to the AI. He knows everything that is going on on the server without being in what we might call viewing distance.
Imagine the surprise on that AI, if it were somehow manifested in the real world. How it would attempt to levitate objects - including itself - around... Eerily funny...
I'm going to disagree with you, Tek Guy, because the researcher clearly states that Micah is gone and can't see what is going on. Regardless of the location, he was wrong by the rules that were set in the test. That being said, this is an amazing jump in artificial life. MANDROIDS FOR EVERYONE!
Great, the people at my old school are spending beaucoups money on not only playing Second Life, but making it even more creepy than it already is. Well done.
Oh em gee!
The machines are going to take over the world!
Tell him that everything you say is a lie.
Hey Eshu,
as far as the AI is concerned, seeing something does not mean knowing something and vice versa. It has no eyes, and thus isn't limited by our concept of seeing. There really is no reason for it to believe that micah does not know what has happened (when in fact he does!), just because it is told that he can't see something. Don't get me wrong, I love the idea of an AI running on a supercomputer to appear as an avatar in something as trivial as SL, but this is little more than a publicity stunt. They could very well have run that test on the AI, but surely not in SL for the reason I just pointed out.