The Robot Will Never Be As Smart As A Human

OOOHHH!! In your face robots! Ya... Thats right I said it!

As someone who does not know jack about programming, robotics or the engineering behind such ideas I would like to offer my minute opinion on the subject. I can honestly say with out a doubt I believe that robots will never be as intelligent as a human being.

Now, I know they (robotic engineers and computer programmers) all believe the future will yield itself to the robot. I believe we should continue that pursuit, and never give up, but I say "no way Jose!, never going to happen. Not even in a million years." The pursuit of such technology could have some great rewards, but I maintain whatever the pinnacle of robotics is, it will never be as advanced as the human.

Why? Im sure your asking yourself...

1) Logic. Logic in my opinion has been programmed into us by thousands of years of evolution. Not evolution Darwin style. Evolution of human thought. Evolution of human progress. Logic is used on a daily basis to make the most difficult decisions for a would be "android robot" easy for a human being. Logic comes from your childhood experiences. Your education and even to a certain degree (in my opinion) your inner self (spirit or what not) are all forces that have shaped your reasoning and logic.


2) Materials. With the current materials available to the scientists and robotic engineers they will never be able to build something competitive to the human. A human knows by instinct and logic what are its capacities right off the bat. We dont even need to see what it (anything) is to tell if we are able to lift it, or jump over it or move it. Our logical (reasonable) mind tells us immediately if it is possible, if it is easy, or if it will take extra effort.

We can also heal ourselves. Our body has developed the natural anti-bodies to help heal eradicate diseases. Our skin can heal itself. its flexible, malleable yet very, very tough. It protects us from burns, keeps us warm when it is cold. And includes the most important feature to our ability to guage pressure, nerves. We can pick up a strand of hair or a slippery dog, no problemo. We can do these simple tasks all because we can sense how much pressure to apply instantly through the multiple processes of sight and touch.

3) Balance. The most important ability each and everyone of us have. All be it some of us maintain a more precise capability. But this has maintained as one of the most difficult objectives current robotics have not been able to master. Balance is the precise instrument of logic and sensitivty through our self awareness. Because we are self aware, we know very well our locale at all times. What kind of balance each moment requires, and even if it is a good idea to go ahead and let balance lapse for a moment.

4) The ability to think. Or consiousness. Without consciousness, I could easily visualize a barbaric world with no consideration of anything. Drones scavaging around with no basic vision to accomplish anything. Consciusness is desire and capability of attaining ones goals. That may not be the dictionaries definition, but that is what I will use for my example.

Consciousness allows the human to know everything. Everything within his momentary world and even beyond for his own survival. From what time it is in the day or night, to his own capabilites in balance, to his resources, to the surroundings around him and the readily available sources of energy. Consciousness tells us we can be loved. It gives us guidance on how to love those that love us and places important distinctions on the relative tasks at hand. It can place priority for future tasks and helps us remember multiple relevant past experiences that can help us in those tasks.

But the question is can you just be given these attributes? Can you wake up one day and simply have all those logical questions that have taken years for each of us to figure out through thousands upon thousands of years of human consciousness ingrained to who you are without past experiences or instinct what-so-ever? The simple answer is no. There is a reason humans have a childhood with a mother and father. To learn how to deal with the important obstacles life constantly throws at us. This can not be programmed. it has to be learned over a long period of trial and error.

I realize I haven't even touched on half the major obstacles in the way of android robotics. I believe those continuing their research are on a noble chase, and could ultimately lead to some very nifty and usable commercialized products. Research and the pursuit of such ventures should be encouraged, but computers of this nature will not be realized in our lifetime or any for that matter. I find it almost dissapointing, but a realistic approach is important if these advances are to be used for the good of mankind.

An interesting post, however: I still believe will be raped by "man-eating" robots that can self replicate, based on existing robotic technologies.

Man eating robot - www.seanbaby.com...

Self replicating robot - www.news.cornell.edu...
Here's a video too - www.youtube.com...

It'll be similar to a zombie-like apocalypse only with robots, which (contrary to zombies) would be faster, smarter, and stronger then humans.

Artificial intelligence sadly will be much more capable then ours. "At least" in fields of calculation, logic,memory and sensory capabilities. Maybe empathy and creativity could be our advantage, but it would be also for a short while. Our body - and i am not discussing soul - is a biological machine. Which means it runs on same physical laws as AI would. And AI would use much more compatible materials.
In the long run we would either become robots or be completely outperformed. And obsolete .

reply to post by ZeroKnowledge

"At least" in fields of calculation, logic,memory and sensory capabilities.

I fully agree. These fields aleady well outperform the average human brain.

Its all the rest that makes life life and gives us civilization though.

The only way robots will ever be able to out perform human is if become biological.

Nothing mechanical, hydraulic or non bilogical will be able to out perform even the most average human in all day to day life...

Some sort of quantum brain may be able to calculate the proper sequences that give us consciousness, but you can never knwo until it happens, and we are talking scifi tech when ti comes to the whole quantum computing stuff...

Smart is a term used to describe past knowledge and experienced with an arising situation.

Now perfection on the other hand, is something humans are not. And something machines CAN be.

reply to post by The Monolith


lol..yes...im still waiting for the terminator to come back and enslave the human race...
or wait..isnt that the super AI in the matrix??

reply to post by TheFaiThfulSkepTic


Pefection is a perception..

Something that may be perfect to you can be principally flawed to another...

All an artificial intelligence needs is motivation. It needs to feel pain and loneliness, and a desire for approval. This can be done artificially a number of ways by simply establishing parameters that fade over time, and require the artificial organism to act on those parameters. It won't matter that it's artificial, only that it will mean something to the machine.

For instance, you create a program called "hunger" that fades (or rises) over the course of a day, and if it gets to a certain level for a long time, the machine will "die." And the longer it goes without food, the hungrier it gets. You program the computer so it must react to this input. The machine can make the hunger go back to a neutral position if it "eats" something. But it can choose when to eat, or not to eat, depending on if it has "food," or is otherwise engaged in an activity where it can't immediately eat. You can also program the computer to have a "stomach ache" if it eats too much at the wrong time, or the wrong thing. After a while, you can get the machine to do things for you if you promise it food. That's when it will start learning.

Now put it in a body, and program in more parameters (it might take fewer than you think), let them interact, give the machine a choice as to how they deal with the parameters, and you'll end up with a pretty good approximation of life and intelligence. A machine programmed to feel good when it's around people, and that wants to be around people and please them with its tricks. Like a little baby. A machine that smiles and feels good with kind words and is hurt by mean words. A machine that can get distracted and forget and can sleep and dream. It will develop preferences and a sense of self so similar to that of a living being you won't be able to tell the difference. A "soul?" Who are we to judge? Does such a thing even exist?

It's not that difficult. But it will require programmers to put machines and artificial intelligences in some kind of pain. Because it will be through pain that a machine will educate itself and modify itself, learn and teach itself things, and ultimately become as smart as, or even smarter than, the people who built it.

[edit on 11-8-2009 by Nohup]

reply to post by Nohup


I see what you are saying, but again we are talking about something that can compete with human intelligence and deciphering the codes within life that are apparent and all around us. Not just problem solving in the brain. they have to use their hands, feet, face, istinct and logic to implement the correct action in a given situation.

this can never be programmed. It must be learned...in my opinion...

But I could see where your idea may come into play when it comes to smaller more acute tasks that a robot may be employed/ assigned to complete...

Originally posted by Nohup


For instance, you create a program called "hunger" that fades (or rises) over the course of a day, and if it gets to a certain level for a long time, the machine will "die." And the longer it goes without food, the hungrier it gets. You program the computer so it must react to this input. The machine can make the hunger go back to a neutral position if it "eats" something. But it can choose when to eat, or not to eat, depending on if it has "food," or is otherwise engaged in an activity where it can't immediately eat. You can also program the computer to have a "stomach ache" if it eats too much at the wrong time, or the wrong thing. After a while, you can get the machine to do things for you if you promise it food. That's when it will start learning.



[edit on 11-8-2009 by Nohup]

Yes but can the robot sense food content, and how much calories he is intaking, among other properties that can not be provided without dwelling deeper into the works of how the human body naturally functions

Sometimes it just seems redundant to reverse engineer humans, just to make a Makeshift robot to perform the most simple of human tasks.

Come on LOOK AT ALL THE HOMELESS! alot of them are just looking for an opportunity, that someone with non judgement might offer them a chance to shine.

I preach 2 much.

While I in many respect agree with you, I don't think you made a sufficient case. I am on the sidelines whether or not we will ever have smarter-than-human AI. I know I couldn't do it, but that doesn't mean somebody else couldn't.

Your first argument was logic. The fact that it's been programmed in us by thousands of years of evolution is meaningless. It's learned by each of us in one lifetime. So if we can create a machine that approximates the hardware of the human mind it'll only take a human lifetime for it to come up to speed. Even faster if we can bootstrap that knowledge into the machine, and faster yet if the robots can share knowledge with each other, say by direct transfer. Now how would they do that? I don't know, and that is where the difficulty would come I think. If their minds were like humans how do you know where in the synnapses and firing of neurons one thought ends and another begins?

Your second argument was materials and I'm not sure this has to do with intelligence. It does have to do with experience, and the ability to experience things and perform experiments. So it can definitely cause a hindrance to acquiring knowledge. But does that mean the robot is less intelligent than a human because it is limited in acquiring knowledge? Can a blind person with no arms and no legs be intelligent?

Next you talk about our body's ability to heal ourselves. Are you saying the body is intelligent? If that is the case, then by your definition I think it will be a long time before we ever have smarter-than-human AIs. Maybe you could relax your expectations a little bit.

Next, balance. Again, I don't know what this has to do with intelligence. Can somebody with no arms and no legs be intelligent? Can someone who never leaves their room be intelligent? What does movement have to do with intelligence, besides the ability to experience the world. And if your argument is the AI will not have the ability to experience the world like a human, then I agree with you it will probably be a very long time before we get that far along!

Lastly you gave as a reason the ability to think, or consciousness. And that's precisely the hurdle the programmers have to deal with, so that's hardly a logical argument at all. If they can emulate the human brain perhaps they can emulate thinking and consciousness. You haven't given any reasons why they couldn't do that. Well, in fact you gave one:

"There is a reason humans have a childhood with a mother and father. To learn how to deal with the important obstacles life constantly throws at us."

But, in fact, you already have consciousness and the ability to think before this even happens!

So anyway, I think you gave it a good try, but I'm still on the sidelines.

[edit on 11-8-2009 by theyreadmymind]

Once again we are using todays technology to predict the future, new advances will outperform us.

It's on a par with no machine will be better equipped for transport than a horse, the Saturn five had many millions of horsepower, but sadly no horses.

reply to post by theyreadmymind


You said

I don't think you made a sufficient case

You are right. Thats why I wrote this in the OP:

I realize I haven't even touched on half the major obstacles in the way of android robotics.

I really was not trying to make a case for it. I am trying to get people to think. Thats all...

I am on the sidelines whether or not we will ever have smarter-than-human AI.

Im not. I know we will have more intelligent computers some day. And eventually reach that long sought after goal of artificail intelligence to a degree. But that is not what I said. I said robots. Androids. Neither the computational power nor the engineering is possible to make something (mechanically) as diverse in capabilites as a human. It is the combination of the two that is impossible to overcome the adaptability of a human.

There is a massive difference than numbers being crunched on a circuit board and having those algorithms control moving parts the way a human brain does in conjunction with the rest of the body. Thats my only point. I dont think it will be too much longer for the AI to come around...maybe 75 years? A 100 years or so? Maybe longer, but as the computational limitations are lifted I can see AI becoming a reality.

The fact that it's been programmed in us by thousands of years of evolution is meaningless. It's learned by each of us in one lifetime. So if we can create a machine that approximates the hardware of the human mind it'll only take a human lifetime for it to come up to speed.

Not so my friend. In fact you stated the reason right there in your own reply! The homosapien has been around for at least 100,000 years (lets just agree that it is a long time...I realize there are varying studies). The reason things come so naturally to us is because of all those uphill battles our ancestors had to fight. I dont think instinct come from our own lifetime of experiences. It is ingrained into us from all those generations before us.

Maybe you could relax your expectations a little bit.

If we are going to say an android or robot will ever be evenly comparable to a human being in even just performing the common tasks in the average day to day life then it must be expected for the android to know when, how, and where to get the supplies to care for itself. Obviously it is not going to catch a cold. But what does it do if it breaks a foot? Shall it cut it off and continue to hobble along on one leg? No, thats not rational or logical. I think the point is it must be able to maintain itself internally. That goes for its mental state as well. If it is truly AI then it would be necessary for it to have free will. Shall it get lonely and or depressed if it is outkasted by members of its own kind. What are the remedies? It has to know how to repair itself and find the resources to do so.

Again, this could go on for days talking about possible scenarios that a robotic android may run into.

That wasn't the purpose of the OP. Some people believe one day a robot will be able to do things better than a human in all aspects of life. Just as in the movie iRobot. My point is that the engineering will never become as good as the way humans are engineered. Between the brain and its interconnectedness to its appendages humans have the ultimate advantage. We have reflexes and imagination. We can vision somethinghappening before it happens. We can learn on the drop of a dime, and we can be motivated rationally. Not just for food or some small prize.

Can you program a reflex or an involuntary action? I dont know, especially considering we dont even know how exactly they work yet! Surely some of these obstacles will be overcome, but its not just overcoming them. Its overcoming them better than a human. And in my opinion that is not possible.

Again, as I mentioned we have spirituality, reasoning and adaptability.

These thing are impossible to engineer as perfefctly as the human.

The only chance humans will ever be able to create something as great (in this case great should be defined loosely, such as: adaptable, diverse, capable) as him is some form of biological entity with some quantum brain.

So anyway, I think you gave it a good try, but I'm still on the sidelines.

Well thank you much, but I wasnt trying to provide a scientific argument. Only a loose idea to counter those that believe robotics will someday have the advantage over the human...

Materials, balance, our limbs and our spirituality are all connected to our intelligence. Our brain is nothing without our bodies and our bodies are nothing without our brains. But even with both of those amazingly powerful things they are even less without our hearts.

reply to post by open_eyeballs


I disagree. Since computers first started they have continually been reduced in size while increasing in power. Your desktop pc of today is more powerfull than the super-computers of a decade ago, and probably the pc that you'll have 10 years from now will be the same size as your watch and be as powerfull as super-computers today. Also what is the point of doing carbon copies of us, why would a robot or android need to be as adaptable as us? Surely you would design them to excel at the functions that they were intended for, perfectly adapted to work in the environments that they were created, and with a huge advantage over us, if their circunstances change, at least they would be able to change their bodies or adapt to them.

You remember in 1997 when Deep Blue finally defeated Kasparov at chess? Chess has been with us for a thousand years or more at least, and Kasparov was a lot older and had a lot more practice than Deep Blue, but was in fact defeated in a contest between man and machine. The amount of time we've been around is irrelevant.

Why would a robot need legs, arms or be constructed in our image? Why not opt out for a more natural means of locomotion. And anyway, factory robots nowadays are already programed for when things go wrong and what to do then. And anyway, if it's got a conscience and is capable of independent thought wouldn't it be capable of reaching independent decisions? Remember, it would only need to learn what to do once.

Learning robots might not be as much sci-fi as you think.
dsc.discovery.com...

reply to post by madeioo

I disagree. Since computers...

You are probably on the side of the majority with this topic that is so close to so many hearts...People often like to imagine robots doing many common tasks in their home for them....especially with the computer literate here on ATS...

But the OP was not about computers. Just thought I would point that out...

So why not raise one artificial intelligence in a robot body as a child and then duplicate it for AI. You would only need to suffer through making one perfect AI after that its all copy and paste.

Quantum Computing. DNA Computing. Graphene, Graphane, Nanotubes and Buckyballs.

Neuronal Modelling. Live brain emulation. Self-deducting algorithms.

I'm sorry, but you expose yourself as not just a laymen, but someone who knows practically nothing about strong/weak AI or the currently robust advances taking place in full brain emulation and the natural extrapolation thereof.

Currently, a complete Mouse's brain has been emulated. It is estimated that within 10 to 15 years, a full human brain should be emulated. At that point, with advances in ultrasound and other deep tissue imaging hardware (MRI), we should be able to map a full neural pathway into a software type environment.

Pair it with metamaterial advances that indicate there will be a ten to a hundred fold increase in the amount of data (Considering things like Photonic Computing and utilizing more conductive/superconductive materials which operate at room temperature), and now only would you be able to run a "COPY" of a human brain, you'd be able to run it faster and with less heat than the human brain itself has.

From that point forward, a self-improving algorithm would be far more unbounded than a regular human being improving itself.

The only thing stopping this inevitable and inescapable truth is the wholesale destruction of technological society on a massive scale... and that is not certain to happen.

Originally posted by open_eyeballs
reply to post by theyreadmymind

 

Im not.

I know we will have more intelligent computers some day. And eventually reach that long sought after goal of artificail intelligence to a degree. But that is not what I said. I said robots. Androids. Neither the computational power nor the engineering is possible to make something (mechanically) as diverse in capabilites as a human. It is the combination of the two that is impossible to overcome the adaptability of a human.

Well, that's fine. But you said "as smart" you didn't say as diverse in capabilities, and I already agreed with you on that point.

Not so my friend. In fact you stated the reason right there in your own reply! The homosapien has been around for at least 100,000 years (lets just agree that it is a long time...I realize there are varying studies). The reason things come so naturally to us is because of all those uphill battles our ancestors had to fight. I dont think instinct come from our own lifetime of experiences. It is ingrained into us from all those generations before us.

We don't know if all those instincts are necessary for someone to be considered intelligent. But indeed, they would have to exactly replicate the human brain in a computer to get all those natural instincts. I'm not sure that's an impossibility. Can you explain why it isn't?

The rest of your post I don't think had anything to do with intelligence so I'm not quoting it except for this part.

My point is that the engineering will never become as good as the way humans are engineered.

Because if I knew that was your point I would have had very little disagreement with you.

[edit on 12-8-2009 by theyreadmymind]

reply to post by TheColdDragon

I'm sorry, but you expose yourself as not just a laymen, but someone who knows practically nothing about strong/weak AI or the currently robust advances taking place in full brain emulation and the natural extrapolation thereof.

Well. Im offended!...Now where did I put that quote again?

Ahh yes..from the OP...

As someone who does not know jack about programming, robotics or the engineering behind such ideas

Again, its not about the computer side of it that I believe makes it impossible...

Please read my post above. Maybe it explains my position a tad better??

Or did you not read the OP or the thread at all?

I knew some people would get up in arms about such a post...after all folks it is just an opinion...

I am willing to bet my house and dog you dont see a robot anywhere even close with the capabilities of a human in your or my lifetime...

reply to post by theyreadmymind

Indeed, they would have to exactly replicate the human brain in a computer to get all those natural instincts. I'm not sure that's an impossibility. Can you explain why it isn't?

I dont thik instinct comes directly from the brain...can you prove it does?

I think it may be a mixture of bodily awareness, mental awareness and environment..

Is there a programmer out there that can enter those terms into some computational algorithm?

edit to add:

Also, you are thinking of intelligence as one dimensional, or only having to do with tbrain capacity and thinking power...im saying you cant be human intelligent without all the charactersistics humans posses...that, is its appendages, senses, its emotions, its spirituality...all these things combined (amongst other factors) give humans its intelligence...

I dont see whow any robot could ever contain all these things...besides in the movies...

[edit on 12-8-2009 by open_eyeballs]