Inspired by biological systems, scientists have developed miniature robots that can self-assemble using parts that float randomly in their environments.
The robots can correct their own mistakes.

AI, robots that build themselves, one rogue robot that watches "Terminator" or "AI" and gets some ideads in it AI....

MSNBC

First of all may I point out that AI robots will never conquer us as films like I:Robot suggest. Robots lack something that humans have, a conciousness. This is what governs us as life forms to do certain things that appear unpredictable or things that make us individuals. A robot can never have a natural conciousness and therefore can't see humans as needing to be destroyed.

As for the mistake correcting process, it is something that many scientists and us web developers have been working on since the thought sprung from the fountain of ideas. To make a system that self operates and learns is the ultimate. Systems like these have been made numerous times in various scripting languages and many games hold on the idea of adaptation. For example, if you play the same move over and over on a game like Tekken 3, the computer will eventually start to learn its mistakes and works out a logical path and beats you. It is the same with any problem.

It is only a matter of time until the AI world evolves into something that co-exists with us as humans. But with so many narrow minds inhabiting this planet, I doubt it will be in the near future.

Interesting theory, let me say you something. Until scientist don't think up some robot that is able to learn by every step it takes, AI will not evolve. When this important step occurs, there will be a great revolution of robots. AI is important part of every robot, but it's programmed, and it's hard to simulate life with program software. When scientists could program only minimal input of AI to a robot, and robot can then start self-learning process - that will be a real AI-robot.

Well the progress of AI is getting more advanced with time, but that a program which would think as a human, for now days is impossible, maybe/really in the future it will be available, besides robots can't have emotions, they can copy them or someone can write a program, but that would be artificial. Well, if the program will understand/make a conclusion by self learning something bad, you can't know how it might act.

Sorry to upset you... But if current trnds continue, we will have computers powerfull enough to simulate the electrical activity of a human brain by 2025.

by 2030, computer will be powerfull enough to also emulate the chemical influences in out brain.

so in 25 years, we will have neural network computers, capable of doing anything a human brain can, if properly taught.

If computers ever turn againsed us, it will be the NEURAL NET which allows them to do so.

Its completely possable...

In conventional computers, like the one you are sat infront of, humand program them,..
so the computer does exactly what the humans programmed it to do.

With Neural networks, NOBODY programs them... They Learn, mimicing the way human brains work.

there is no reason that a neural network could not be everything humand are, just made of different materials...

Read up on Neural networks that can drive pretend cars in computer games...

You dont program them to drive... you just teach them that they can avoid objects by turning, and/or slowing down...

they learn, they program themselves... and then they drive around a mase, completely un-predictable..

they *CHOOSE* which way to drive.

Neural Networks are Fascinating things.

Very interesting stuff here. I was amazed to see the robots dancing in synch. It's just amazing at how much a robot can accomplish, and what they can actually do. I kinda agree however, that robots lacking a consciousness will prove that we humans are far superior Or at least I hope so. No telling what the future holds.

Computers will never be powerfull enough to simulate the human brain. Even if they do have all the mechanics of the brain. Why? Because we haven't even scratched the surface of what our brain is and how it works. And I don't know where you got your imformation from, but I would like a link to see it.

As most of us know, we use a small percent of our brain, but robot brain can make very fast decisions calculating all of the abilities and making the best conclusion, which you would be thinking for days and etc. This is something superior, or a better feature in my opinion.

Scientists have often said that we only use 20% or so of our total brain power. If we were to use all of it, we would advance quicker. But I believe that we don't use all of it as we are using too much thought power for problem solving. My proof, we are constantly problem solving. From every step to decisions on whether to eat or not.

Anyway, the human brain also brings up numerous of other results, as well as the one that the robot brings up. I am love technology, but something that I can't see happening is robots developing a sense that we as humans have. They may be programmed to learn and adapt, but they will never have a natural conscience.

well, now it looks imposible, but actually self learning robot designs already exist, of coruse they are very primitive, but it is the start.