Can machines become self aware?

[MasterHamster]

Why is a mechanical form of your neurological patterns not as sentient as yourself?

Still based on algorithms and thus an illusion of sentience.

No, the person will produce an answer based on their past experience. So can an AI, based on the information it has received in the past.

Except a machine cannot imagine a scenario without set patterns on how to do so :3

Just as you are biologically programmed to be averse to pain.

Pain it didn't experience, because it has no mind
Yet another imitation

[Heathy]

like ppl have said it doesn't matter how much programming you do, it is still programmed so its not self aware its still running on commands based on its current situation.

could you imagine how much programming it would involve to get a robot to interact with the majority of things in your house, insurmountable, now imagine how much programming it would take to get a robot to interact with the things in your house based on its current circumstance. if you can even program current circumstance. you could write lines of code your entire life and not even get close to an AI capable of reasoning and emotion let alone desires and the ability to perform actions based on what is needed at any particular time. the best i've seen so far is a robot walk across a stage and up/down a few steps. robotics and AI are no where near self awareness, there isn't even any good AI in gaming, the same AI exists in most games as has existed for years, it never does anything sporadic or 'for the lols' or because it might work, it doesn't make mistakes it doesn't learn from mistakes it mostly just takes the best possible option that it can based on calculations of the multiple outcomes from the options it could take, RTS AI is a good example, because RTS gaming has more to it that run around shooting AI, thats easier to accomplish, getting a computer to actually pull off tactics is nearly impossible (unless its already been programmed to do so) it never does anything you wont expect actually most AI in games cheats and cheats a lot to have even the slightest chance vs real intelligence.

robots don't eat or sleep or even have the ability to enjoy something it has no desire or agenda, no moral compass no ability to care or show empathy or pity or envy etc etc, so the question remains, why would you create a work horse with self awareness that would probably only end up learning its a slave.

i think we need to define self aware, to me the definition means that it need to be aware of its self existence, which includes feelings and emotion toward that knowledge also free will as loosely as it can be, some disagree free will even exists, i am on the side that says it does exist and choices aren't always made for the best reasons or the most benefiting reasons or for any reason. i think self aware and free will are intertwined, you can't have one without the other.

in the end it really does just come down to any machine will only be as smart as the programming its run on, even if it learns it will still only learn from what it can't accomplish in the first place, it still doesn't make it self aware. just because it can differentiate between succeeding and failing.

[Deleted]

No, the person will produce an answer based on their past experience. So can an AI, based on the information it has received in the past.

The guy doesn't seem to know that your brain does the same processes of a computer using your experiences as the input information.

This does not prove sentience. What does is how the person or AI will use said infomation.

[Dragonix80]

This discussion is impossible.

On one hand we have those who regard AI as a construct, a machine that is told how to process information, in order to "learn". No matter how advanced, it'll remain a heap of scrap that uses enough algorithms to appear sentient.

On the other, we have those who think mind, sentience, self-awareness and sapience can all be created by making a machine that simulates neurological patterns well enough.

Well I'm not buying it. While I am a person that thinks that in the end, we're just a bunch of cells, chemicals and neurons, a true, living mind cannot be constructed just by imitating well enough.

Ask a person what he thinks will happen to X when Y happens, he'll consider an answer from personal feelings, bias, ignorance, knowledge and experience.
A machine could come up with an answer after running a few quadrillion simulations and deducting the most likely outcome.
It's a goddamn illusion and nothing else.

I agree with your statement.

You don't know, you only think you know. If you come from a religious stance, it doesn't matter if you say you're leaving it aside. You cannot truly leave it aside, and we can never communicate. Mine is a mechanical and scientific universe, there is no magic or miracle here.

Where did you come off with the very idea that whatever I believe in is not mechanical and scientific but magic or miracle? On the contrary, science have strength my faith in higher intelligence or maker (Apparently, I'm forbidden to say the dangerous word because Scrapbot is programmed to seek and hunt anyone who say the forbidden three letter word.) So no. I do NOT believe that there will ever be a true AI machine, only machine as MasterHamster say, "No matter how advanced, it'll remain a heap of scrap that uses enough algorithms to appear sentient."

[Deleted]

Still based on algorithms and thus an illusion of sentience.



Except a machine cannot imagine a scenario without set patterns on how to do so :3



Pain it didn't experience, because it has no mind
Yet another imitation

All of these are things your brain does exactly how you described. In Theory, are we all not self aware then?

[XDurionX]

Of course they can, sufficient complexity provided. We´re nothing but biochemical machines ourselves. And free will is an illusion.

[MasterHamster]

The guy doesn't seem to know that your brain does the same processes of a computer using your experiences as the input information.

This does not prove sentience. What does is how the person or AI will use said infomation.

"The guy".

AI is limited to the patterns we give it. That's all there is.

[Deleted]

"The guy".

AI is limited to the patterns we give it. That's all there is.

Human Intelligence is Limited to the Patterns nature gave it. That never stopped us from being Self Aware.

[semaphore]

Still based on algorithms and thus an illusion of sentience.

How would it be based on algorithms? It's a replica of yourself. The individual components interlink and work together to produce thought.

Except a machine cannot imagine a scenario without set patterns on how to do so :3

Not yet, perhaps.

Pain it didn't experience, because it has no mind

It received the input from sensors, it experienced it. Just like you do.

---------- Post added 2013-02-24 at 02:44 PM ----------

AI is limited to the patterns we give it.

You're limited to the patterns supported by your brain. Just like an AI is limited to the patterns provided for by its architecture.

[theWocky]

I think eventually a full robotic / chemical being will come into existence from a non-biological source. It's creator will probably be partly biological, however,

First, I think we will see a hybrid fusion of biological and machine. It's only a matter of time before it is easier to fuse man's brain and the rest machine together to enhance humans rather than try to make robots smarter. Get an improved species to make AI better is my opinion.

Soon as we extend the life-span and cognitive ability in just one hybrid like this, man will initially be master and then slave to the new species.

This new species will make a complete new mechanical being with an immortal electronic brain. Maybe even develop the tech to transfer man to machine.

[Deleted]

Well, they can always simulate simple lines of what to respond to certain questions or keywords. But you also have to think about that most actual living creatures on this planet are not selfaware. I might be wrong but from what I can remember it's only Humans, Elephants, Dolphins and Apes that is actually selfaware, or atleast that is scientificly confirmed.

And I wonder if they actually even know how to re-create self-awareness or even if they even know how we are self-aware.

[XDurionX]

You guys know that imperative programming is not the only way? Probability heuristics, deductive inference mechanisms, "genetic algorithms" machines today are already capable of learning the same way we do. The only difference is the complexity of our decisions. Denying the ability to become self aware is nothing but chauvinism. Just because we´re not smart enough to understand how our brain works (yet), doesn´t mean its not possible to imitate its behaviour.

[Deleted]

Well, they can always simulate simple lines of what to respond to certain questions or keywords. But you also have to think about that most actual living creatures on this planet are not selfaware. I might be wrong but from what I can remember it's only Humans, Elephants, Dolphins and Apes that is actually selfaware, or atleast that is scientificly confirmed.

And I wonder if they actually even know how to re-create self-awareness or even if they even know how we are self-aware.

Domestic Dogs and Cats have evolved Self Awareness as a social thing to humans.

[MasterHamster]

Imagine a robot that has a perfect replica of the human brain and neurological patterns.

It could see, but what it sees would be interpreted using algorithms we gave it, such as pattern recognition.
It could hear, but what it hears would be from identifying parts of words and then comparing to a database.
It could feel (like "pain"), but what it feels would be from choosing an appropriate action based on a set pattern.
It could imagine, but what it imagines would be from a set pattern that determines what to include in this "imagination"
It could have urges, but what it wants would be determined from a set of patterns that would be most beneficial
It could have the ability to learn, but it learns from a set pattern on how to deduce information from events and store it.

Of course I can see the similarity to this and the human mind. However, the major factor that leaves this amazing machine as nothing but a simulation is that we have to set every single pattern. It could come up with some patterns on itself by "learning", but that would still rely on us. How could it know how to "learn" unless we tell it how.

It's been fun, but frankly this discussion is far too reliant on "what if" and "how is that any different from [x]".
Make a complete artificial brain, and you'd still have nothing but a heap of scrap unless you tell it how to process anything. It's not going to write the first line of code by itself.

In the end, it's still just an imitation of mind, sadly.

[Deleted]

Imagine a robot that has a perfect replica of the human brain and neurological patterns.

It could see, but what it sees would be interpreted using algorithms we gave it, such as pattern recognition.
It could hear, but what it hears would be from identifying parts of words and then comparing to a database.
It could feel (like "pain"), but what it feels would be from choosing an appropriate action based on a set pattern.
It could imagine, but what it imagines would be from a set pattern that determines what to include in this "imagination"
It could have urges, but what it wants would be determined from a set of patterns that would be most beneficial
It could have the ability to learn, but it learns from a set pattern on how to deduce information from events and store it.

Of course I can see the similarity to this and the human mind. However, the major factor that leaves this amazing machine as nothing but a simulation is that we have to set every single pattern. It could come up with some patterns on itself by "learning", but that would still rely on us. How could it know how to "learn" unless we tell it how.

It's been fun, but frankly this discussion is far too reliant on "what if" and "how is that any different from [x]"
Make a complete artificial brain, and you'd still have nothing but a heap of scrap unless you tell it how to process anything. It's not going to write the first line of code by itself.

What you have there are the basics of Primitive True AI. Those "Parameters" you say are the variables that can make it self aware or not. It just takes one question formulated from all the data it stored for the AI to become properly sentient.

"Am I Alive"

Once we here those words spoken or written by an AI on it's own accord using it's own judgement and thinking skills, it has become Sentient.

[semaphore]

It could see, but what it sees would be interpreted using algorithms we gave it, such as pattern recognition.

Or it could develop one based on learning.

It could hear, but what it hears would be from identifying parts of words and then comparing to a database.

Or it could collect one based on learning.

It could feel (like "pain"), but what it feels would be from choosing an appropriate action based on a set pattern.

Or it could choose one based on learning.

It could have the ability to learn, but it learns from a set pattern on how to deduce information from events and store it.

Look up how machine learning works.

It's been fun, but frankly this discussion is far too reliant on "what if" and "how is that any different from [x]".

If only it weren't so full of people who insists on pretending that AI = programmed algorithms.

[Collegeguy]

It received the input from sensors, it experienced it. Just like you do.

Not so fast. Any computer we can currently make has absolutely no efficient way to record "experiences" in large amounts and store them in a manageable way. The biological brain has an very efficient process for this and inadvertently chooses what data is stored and what data is just forgotten and let go. A computer can't do that, so it stores everything and eventually exceeds its own limitations.

Furthermore, I believe the biological brain can still processes data at a much higher speed than a computer.

[semaphore]

Not so fast. Any computer we can currently make has absolutely no efficient way to record "experiences" in large amounts and store them in a manageable way.

Irrelevant. We're talking about theoretically possibilities in the indefinite future. I don't think anyone disputes that we don't have true artificial intelligence yet. Except this one person who claims to have created it.

A computer can't do that, so it stores everything and eventually exceeds its own limitations.

Why does it need to store everything?

[MasterHamster]

Or it could develop one based on learning.
Or it could collect one based on learning.
Or it could choose one based on learning.

Learning that we have to program for it.

Look up how machine learning works.

Kinda like that machine that was meant to learn how to walk without movement-based code.
Flailing its legs around mindlessly until it recognized that it was moving forward.

An "urge" that we told it to have.


There's your intelligent learning robot

If only it weren't so full of people who insists on pretending that AI = programmed algorithms.

So much hostility.
And for the record, without algorithms an AI wouldn't do anything.

Peace ..\/,

[Dezerte]

Relevant http://www.yale.edu/graduateschool/p...awareness.html

I'd keep my eyes on that robot.