Will Computers Take Over Earth? Computers taking over the world

[Maggot 1]

I see quite a few people asserting that computers cannot be creative, or that AI systems are limited by original design. Not so in either case. The brain is a network of connections and junctions. There is nothing in principle which says that we cannot build an analogous system in silicon (or whatever material). Those who believe there is something 'extra' about the brain - either a 'divine spark', a 'soul' or some other intangible have not made a convincing case, since such views rely entirely on faith. Given that it is possible in principle to build a silicon brain then I see no reason why we would not, eventually, do so.
For those who think AI systems are limited by design - consider a neural network of 10^40 'neurones' and (say) 6 path internconnectedness with full weighting for each in (say) 8 bit code.
How many possible states can this system have? (Answer - a number so big that it is a waste of time to calculate it). Hook your neural net up to some sensors and some output devices and start teaching it. What are the limits of such a system? Theoretically we could calculate an information limit (just as we can for the brain), but it would be effectively infinite (unlike the brain which has much less capacity).
Could such a system be 'creative'? The answer to that depends on whether you think the human brain can be creative. The answer is the same in both cases. You would, of course, have to supply an initial goal or aim. In humans that is what we get in the hardwiring - mainly the circuitry in the brain before birth.

Now, should we be worried? I think not. Why would we build a robot of immense potential intelligence and then make it look like a human? I can see no good reasons, unless it was to mimick a human and humans are much easier to make than robots. Robots and autonoma tend to be made in a form suitable for the function performed. OK we have a few like Asimo that are designed to look human(ish) but they are mainly for show and to test ideas. I don't worry at all about this - though I will be dead before we get systems as intelligent as a human I suspect (ie it won't likely happen in the next 30 years or so).

A computer cannot make decisions for them selves plz.......

If your saying that this is possible in the near future then.... Evry computer now and in the future will depend on a human .....

Computers can never be self sustaining its just a dream like god or the devil......

We can make descisions according to what we think

Computers are always going to follow all the instructions we give them............ Maybe im wrong and computers will take over the world... lol...

[Bikerman 2]

A computer cannot make decisions for them selves plz.......

They can and do routinely.

If your saying that this is possible in the near future then.... Evry computer now and in the future will depend on a human .....

No I am saying that this has been the case for years. Nobody told Deep Blue what moves to make when it beat Kasparov. Who could have done, since Kasparov is the best chess player in the world.Neural Networks (computers that can learn) are used in science, business, medicine and other fields. The famous 4 colour theorem was solved by a computer.

Computers can never be self sustaining its just a dream like god or the devil......

It depends what you mean by self-sustaining. Robots exist which can seek out power points and plug themselves in, as one example.

Computers are always going to follow all the instructions we give them............ Maybe im wrong and computers will take over the world... lol...

You are wrong and computers have been running without instructions for at least 15 years.
Here, go and educate yourself
http://forums.xisto.com/no_longer_exists/

[zanzibarjones 1]

Nope never. Computers right now are stupid compared to humans. In order for them to even come close it would take decades more research. And even then I am most sure they will program some sort of fail safe into them.

[Vistz 1]

I doubt that this sort of thing will happen. If, by some chance, computers do manage to "take over the world", it will be a person that is responsible for programming said computers. "Computers are only as smart as the people who program them." -Scooter Computer and Mr. Chips

[Bikerman 2]

I doubt that this sort of thing will happen. If, by some chance, computers do manage to "take over the world", it will be a person that is responsible for programming said computers.
"Computers are only as smart as the people who program them."
-Scooter Computer and Mr. Chips

The trouble with quotes in computing is that they are out of date by the time you use them.
Certainly computers are dumb by our standards at the moment. Estimates put it somewhere between a bee and a small rodent.
The idea that computers only do what they are programmed to do is just wrong. MOST computers do, but not all. Quite a few systems now use Neural Networks or similar Artificial Intelligence technology. A neural network is not programmed like a 'normal' computer - it learns.
Here is a simple example of a neural network system : http://www.20q.net/
The neural network will attempt to guess what you are thinking of. Nobody has programmed it to do this - it learns.
Now this is a trivial example, meant largely for amusement. Much more powerful versions already exist in industrial and research organisations and there is no theoretical limit to how much they can learn.
Some of my colleagues predict that technology will reach the intelligence of an average human (IQ 100) in about 20 years. I think that is a bit optimistic and would put it at more like 50. It WILL happen though..

[networker 0]

I can see some scary things happening with the combination of Robotics/Digital/Nano
technology.
I think it's very feasible that computers could 'run' the world. All
one has to do is look at what already happened. Years ago offices had
many,many file clerks moving files on rolling carts from office to office. Computer technology has now eliminated the use of these people completely.

That was just a beginning. Of course now they use Robots for everything from bomb
testing to factory automation. We haven't yet seen it even come close to it's
natural conclusion.

The trouble with quotes in computing is that they are out of date by the time you use them.Certainly computers are dumb by our standards at the moment. Estimates put it somewhere between a bee and a small rodent.
The idea that computers only do what they are programmed to do is just wrong. MOST computers do, but not all. Quite a few systems now use Neural Networks or similar Artificial Intelligence technology. A neural network is not programmed like a 'normal' computer - it learns.
Here is a simple example of a neural network system : http://www.20q.net/
The neural network will attempt to guess what you are thinking of. Nobody has programmed it to do this - it learns.
Now this is a trivial example, meant largely for amusement. Much more powerful versions already exist in industrial and research organisations and there is no theoretical limit to how much they can learn.
Some of my colleagues predict that technology will reach the intelligence of an average human (IQ 100) in about 20 years. I think that is a bit optimistic and would put it at more like 50. It WILL happen though..

[mahesh2k 0]

I can see some scary things happening with the combination of Robotics/Digital/Nanotechnology.
I think it's very feasible that computers could 'run' the world. All
one has to do is look at what already happened. Years ago offices had
many,many file clerks moving files on rolling carts from office to office. Computer technology has now eliminated the use of these people completely.

That was just a beginning. Of course now they use Robots for everything from bomb
testing to factory automation. We haven't yet seen it even come close to it's
natural conclusion.

Think about it if they can come to feedback or make decision on their own. Even if they manage it then it is huge learning curve to take over rest of the machines. This will require manual programming to all machines. It's not so easy and will take more than 50 years to get such computers into working. I see this possibility in very small chance. Computers will remain dumb for next 5-10 years for sure.

[Bikerman 2]

Think about it if they can come to feedback or make decision on their own. Even if they manage it then it is huge learning curve to take over rest of the machines. This will require manual programming to all machines. It's not so easy and will take more than 50 years to get such computers into working. I see this possibility in very small chance. Computers will remain dumb for next 5-10 years for sure.

For sure? I wonder how you have such certainty. Are you an expert in AI systems? I'm not an expert, but I know a bit about it (my first degree is in Computing and I did some work on AI systems, but that was a long time ago) and I'm not certain. I know that it would be very easy for a conscious machine to take over other non-conscious machines - they speak the same fundamental binary languages, so once the 'flavour' of processor is known it would be trivial for an intelligent computer to control other digital devices.

[mahesh2k 0]

For sure? I wonder how you have such certainty. Are you an expert in AI systems? I'm not an expert, but I know a bit about it (my first degree is in Computing and I did some work on AI systems, but that was a long time ago) and I'm not certain.

My degree is in electronics and communication and that made me aware of certain limitations of machines/programming. I'm not expert in AI either. My major was VLSI/Embedded system from which i'm deviating for last 3 years. But i still have access to people who are in this field(AI) and i hear few limitations from them often.

My understanding(from brainstorm sessions with others) so far about current state of processor is that, the limit that our single core processor reached and that forced us to use multi core processors for our current working. If any single machine tries to take over another machine will require large processing power to operate itself and then to communicate or manipulate other machines. Like this you can even come up with limitations of quantum processors (which are still in research phase).

The problem with manipulating AI is this -

1. Feedback and Decisions- these machines need to understand human behavior in order to restrict and control them. Which is not possible for AI programmers atleast for next 5 years. Can we see through the spam patterns of people effectively ? without any false positive trigger ? NO. If humans can't be sure of many such behaviors then taking over humans life with some random or ordered set of rules then atleast not possible for next 5 years.

2. Human control - There are plenty of areas where we still hold control and don't automate things. For example many power stations are not connected to internet(atleast no in asia) so there is no way a normal machine can take over power station(just by brute force or something like they showed skynet).

3. Platform - There are variety of OS/Processors/Software manufacturers which use different tools and methods for their internal operation. If one machine wants to take over another using known typical network protocols then that machines has to go through software>OS>processor tree,which is a huge programming work. If any machine programmed enough to take over single platform(instead of multiple) in next 5 years then i could leave this argument.(I'm not talking about virus/hacking stuff, it's about AI automation).

I know that it would be very easy for a conscious machine to take over other non-conscious machines - they speak the same fundamental binary languages, so once the 'flavour' of processor is known it would be trivial for an intelligent computer to control other digital devices.

You have point there. Okay you may argue something like You and I own kindle and our personal information is recorded by amazon to advertise specific books/apps to us based on our reading or surfing patterns. Similarly wowwee robots can be programmed to connect to internet if there is any wifi spot available around to download vendor instructions and then override the programming etc. But these things are still in theory. But i still don't see it as a threat to humans because we still have some control. Not atleast in next 5 years.

[Bikerman 2]

You have point there. Okay you may argue something like You and I own kindle and our personal information is recorded by amazon to advertise specific books/apps to us based on our reading or surfing patterns. Similarly wowwee robots can be programmed to connect to internet if there is any wifi spot available around to download vendor instructions and then override the programming etc. But these things are still in theory. But i still don't see it as a threat to humans because we still have some control. Not atleast in next 5 years.

Oh certainly not in the next 5 years. Maybe in 20, maybe never. It seems to me that as we develop intelligent machines we should also be developing methods of controlling them. It would be criminal stupidity not to....

[mahesh2k 0]

Oh certainly not in the next 5 years. Maybe in 20, maybe never. It seems to me that as we develop intelligent machines we should also be developing methods of controlling them. It would be criminal stupidity not to....

By the way have you checked ALICE chat robot. You can see that robot in action in some messengers like Nimbuzz and in some flash based human chatbot. Improving that chatbot could give some results that we want for example, if anyone wants to talk with someone on one typical topic and if this bot is program to ask and answer question in loop for that topic then it could be beneficial for business, and those who are always in search of person to chat.

[Bikerman 2]

By the way have you checked ALICE chat robot. You can see that robot in action in some messengers like Nimbuzz and in some flash based human chatbot. Improving that chatbot could give some results that we want for example, if anyone wants to talk with someone on one typical topic and if this bot is program to ask and answer question in loop for that topic then it could be beneficial for business, and those who are always in search of person to chat.

Yes - I used it in a lecture with my BTEC students last year. I was around when the original program of this type was written. It was called Eliza and it pretended to be a psychiatrist. It persuaded quite a few people that there was a real person at the other end of the keyboard :-)
http://en.wikipedia.org/wiki/ELIZA

[Harlot 2]

No, I don't think computers or robots will take over the world, and maybe I am being a little short sighted, but I just can't see that happening. A computer is an object, a machine, not a human being. It is not able to sustain itself, adopt to its environment, or make decisions that are in its best interest. Computers are not greedy, deceitful, and they have no secret agendas. They are just machines, a creation that has not conscience, no true thought process, no personality or emotions. Even if a computer has the ability to show emotions, it is not natural emotion, but rather a fabrication. It doesn't really feel, it simply knows how to make a look that represents a feeling under certain circumstances. However, the most important aspect involved for computers being able to take over the world is a true conscience and an ability to adopt to both natural and social change. If it is unable to do those two things, then a computer is dull. It is just another machine, and just as it was programmed to dominate, it can be programmed to surrender. It would have to be a machine with the conscience, reason, and adaptation ability to not be reprogrammed unless that reprogramming is in its best interest. It would have to clearly know and understand its best interest, and its best interest must not be able to be reprogrammed into something that is not in its best interest, unless it believes that a change in interest is its best interest. What programming will govern that belief, and if it can be reprogrammed their dominance of course falls apart. It seems quite impossible for robots to dominate based on the idea that they can always be reprogrammed, and if they can't be reprogrammed...or reprogram themselves, then they are dull and won't survive as conditions change anyway. When you have an artificial intelligence rather than a biological intelligence, it is easy to change that intelligence into something else. You can just pull the plug on the robot, or reprogram it to do something else. When it comes to biological intelligence, reprogramming a human brain is possible but not controllable. You can use media propaganda, and that doesn't work on all biological brains. Then if propaganda doesn't work, you can expose that person to a lot of stress or unbelievable pain and brutality, but what will be left of that individuals brain and thought process can't be predicted. Now, as far as computers playing a large role in our society, that can not be denied. Just as the steam engine started the industrial revolution, the computer processor has started a revolution. A few years ago, you were not able to get on a computer and instantly communicate with anyone in the world, or broadcast live video over the internet with a $20 webcam. More and more companies are using the internet, and so are political movements and government. You have the military becoming more and more intertwined with computers. However, as for taking over earth, I don't think so. Now I may be wrong, and I may be short sighed or closed minded, but I just don't see it happening. However, even if it did happen, would it really matter? All humans live and die, so if humanity ends tomorrow is it relevant if death has to come to all anyway. To live and die...or to never exist....is there really much of a difference based on our current knowledge of the world and universe?

[Bikerman 2]

Well,firstly would you accept that the brain is where we do our thinking, feeling, consciousness etc? Or are you someone who thinks there is some mysterious hidden part of the body responsible for personality, memory etc (they do exist, so I have to ask).Presuming you are sane and you agree that the brain is the place, then consider. The brain is a collection of switches. It contains billions of special cells (neurones) linked together in a big matrix. That is what we do our thinking with.Now, tell me, what is there about this that is impossible to repeat in a machine? Why can't we build a big switch matrix, just like the brain? And if we do then why should IT not be conscious and intelligent?Personally I can see no reason at all. At the moment we are limited - our switches are 2 dimensional - they only switch horizontally. The brain switches in three dimensions which makes it many many times more powerful. There is no reason why we cannot build switch matrixes like that though - and we are.Now, I know some people believe there is a divine spark or 'spirit' which makes us different. Personally I do not believe that, I think it is baloney. So as far as i can tell there is absolutely no reason, in principle, why a computer cannot be as intelligent, as self-aware and as 'conscious' as a human - and more.

[mahesh2k 0]

Yes - I used it in a lecture with my BTEC students last year. I was around when the original program of this type was written. It was called Eliza and it pretended to be a psychiatrist. It persuaded quite a few people that there was a real person at the other end of the keyboard :-)http://en.wikipedia.org/wiki/ELIZA

Cool.

I can't help but think of these BOT app as future way to teach or guide to people on particular domain. I mean improving these bots can help replace help files or guides that comes with software. Just install the info in pre-installed OS bot and let it guide you for working with that software. Just idea in theory maybe in future there will be similar implementation for this. Bozilbuddy was first software which used to do these things and got taken down by court because it was spyware that used to record surfing patterns of users. Still i like that software wish google or some other software giants can come up with harmless app on this concept.