![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
"Researchers led by the Institute of Cognitive Science and Technology in Italy are developing robots that evolve their own language, bypassing the limits of imposing human rule-based communication."
Basically, this article talks about a team of researchers who have taught two of Sony's little "robot dogs" a new trick. They haven't taught the critters how to talk to one another -- they've taught them how to learn how to talk to one another, among several other things. The example they give in the article is that one "dog" can see a ball, and can tell another "dog" where the ball is, if it’s moving and what colour it is, and the other is capable of recognising it as the same ball.
From a technical and Artificial Intelligence standpoint, this is FANTASTIC! It represents a huge breakthrough that can open up all sorts of things.
From a human standpoint, I find it a little worrisome. Artificially-intelligent robots that develop horrible problems are a long-standing science-fiction cliche. While I doubt we have much to fear from roving packs of Rampaging Killer AIBO-turned-Cujo bots, I know techies too well to think that they'll limit these new techniques to AIBOs.
I don't want to sound like a reactionary techno-phobe who jumps at digital shadows. I am all in favor of advances in robotics and in artificial intelligence. However, I've done enough work with A.I. programming to know that as soon as you give something the capacity to learn, you lose control over what it will learn, even when you think you're keeping a close eye on it. Personally, I hope that any machine that has a learning capacity also has some sort of "kill switch" built into it -- one that can be easily used by anyone, and that cannot be disabled.
Very interesting... but I really hope they keep a very close eye on where they're going with this, and I hope they're wise enough to build in LOTS of safeties.
Basically, this article talks about a team of researchers who have taught two of Sony's little "robot dogs" a new trick. They haven't taught the critters how to talk to one another -- they've taught them how to learn how to talk to one another, among several other things. The example they give in the article is that one "dog" can see a ball, and can tell another "dog" where the ball is, if it’s moving and what colour it is, and the other is capable of recognising it as the same ball.
From a technical and Artificial Intelligence standpoint, this is FANTASTIC! It represents a huge breakthrough that can open up all sorts of things.
From a human standpoint, I find it a little worrisome. Artificially-intelligent robots that develop horrible problems are a long-standing science-fiction cliche. While I doubt we have much to fear from roving packs of Rampaging Killer AIBO-turned-Cujo bots, I know techies too well to think that they'll limit these new techniques to AIBOs.
I don't want to sound like a reactionary techno-phobe who jumps at digital shadows. I am all in favor of advances in robotics and in artificial intelligence. However, I've done enough work with A.I. programming to know that as soon as you give something the capacity to learn, you lose control over what it will learn, even when you think you're keeping a close eye on it. Personally, I hope that any machine that has a learning capacity also has some sort of "kill switch" built into it -- one that can be easily used by anyone, and that cannot be disabled.
Very interesting... but I really hope they keep a very close eye on where they're going with this, and I hope they're wise enough to build in LOTS of safeties.
no subject
~
no subject
~