Can machines become self aware?

[BatteredRose]

And now I want to watch AI.

[semaphore]

Sure, why not.

And if Moore's law keeps going at the rate that is going, we are very close to reaching a moment when the computer becomes more intelligent than a human being.

Lol, not even close. Processing power =/= intelligence.

However, there is an evolutionary phenomenon occuring with a new 'species' of human emerging called homo evolutis.

Yeah a new species. Or a book title.

Predictions do say by 2020 we will see a sudden change in the so far smooth exponential curve

If by "prediction" you mean "speculation", and "by 2020" you mean "arbitrarily invented date".

---------- Post added 2013-02-24 at 10:11 AM ----------

The real question is whether we are self aware.

You failed the Turing Test?

[Ifeanychukwu]

I think that given enough time technology can do almost anything.

[Asphyxiate9]

I dont think they could be fully self aware in such a way where they program them selves as they go. Even if it did get to that stage I imagine the technology required wouldn't be thought of for a few thousand years; and once it is it will take even longer to scale them down like silicone chips.

[Banzhe]

The real question is whether we are self aware.

If we were, it wouldn't be each nation to it's own fighting for their "own" interests.., if anything I'd say machines should be rushed with AI in mind, perhaps when they evolve, we become aware.

[Heathy]

i dont think so, machines have to be programmed, even if they some day create one of them movie AI's that answer questions with a degree of understanding, in reality an AI as such would still only be able to recite replies from a pre-programmed list of possible answers.

a better question is if cyborgs existed how much flesh could you replace with machinery until it was more robot than man. or would there even be a man left after say a 70% replacement.

i guess you could program a robot to mimic self awareness, as in, completely freeing up its ability to execute any command it has at any time, the problem with the whole idea is that the amount of programming involved in AI is pretty vast, if you had a robot that was given commands such as walk forward, backward, side step etc etc, the freedom to move its arms and legs at any particular time.. if there is no templates for normal behaviour what you'd probably end up with is a robot executing random commands at random intervals which would probably end up with your robot on the floor flailing its arms and legs around. whatever a machine does has to be programmed into it first, so no its very unlikely that machines could think for themselves not without some extremely random behaviour.

[Deleted]

Of course they can, havnt you seen that documentary on the tv whats it called again? Oh yes Terminator. It tells you all about it.

[galluk]

They will. they are already working with machine that learn from their mistakes. SO I think it is coming. maybe near the end of my lifetime, maybe 40-50 years from now.

The real question is whether we are self aware.

hard to say if we are Self Aware. would venture to say we are more Self Delusional, which makes us think we are self aware.

[klaps_05]

from a theoretical point of view, programming computers/robots to be self aware is inevitable as once programmed they can't make mistakes as long as the system functions.

in addition a computer/robot could work indefinitely, thus rendering it much superior to a human.

think of it this way - if programs/robots could be self aware or programmed as such, you can simply send endless amounts of them to other planets and make them colonize by creating suitable for humans environment.

[Deleted]

Wow I did a project on this for my computer science degree, basically, its debatable basically but I think first we'll eventually be able to create a machine that will appear self aware aka have so many programmed functions that it'll seem like it is and that'll happen a LOT sooner than we'll be able to create one that will truly become self aware, just gotta remember the three laws of robotics when we get there

[Spl4sh3r]

The real question is whether we are self aware.

Actually the question should be how do you define self awareness? Since we can't know if we are self aware or not unless we have a fixed defination of it

[Dreamless]

A self aware program has already been made. Its awareness was limited, but existent.

This was an experiment involving a virtual rat, given only very basic programming in terms of things it 'liked' and things which 'frightened' it, as well as some basic instruction in how to control its body. Fear of water was one of these basic commands.

It was then released in to a virtual tank of water with platforms under the surface and observed. The first few times it 'swam' around in a good approximation of panic, until it stumbled across one of the platforms. On successive runs it got better and better at finding the platforms, until it could get right to them.

Consciousness, from everything we know, is an emergent property of the brain. Any sufficiently advanced system will demonstrate this emergent property to one degree or another; making sentient machines effectively an inevitability as complexity increases.

They won't think the exact same was as humans do, but they will be modeled on us, the same way that children tend to model their morality and ethics on their parents.

[Deleted]

It's not impossible and I'd go so far as to say it's inevitable. We're not so different ourselves, our brains are just vastly more complex and the billions of individual pieces are difficult to replicate at this time. We can make technology that can learn from other pieces of technology already, they can allocate unknown words to objects and by the end of the day know a small language we have no understanding of. The problem is just the number of command strings required, we don't have a way to get these messages across and store them like we do with a brain and even ours is imperfect when you look at the sheer number of mental illnesses we have.

[Wolfbear]

Machines like the big dog freak people out, so they are very tough to implement into the teal world.

It freaks people out because it looks like a Gekko from Metal Gear Solid.
http://mrsthordarson.edublogs.org/fi...ow-1rwuzbb.jpg
http://images4.wikia.nocookie.net/__...d/GeckoArt.png

[Kuntantee]

You got AI wrong. It's not true intelligence. It's simulation of behaviour and you program it. Machines are stupid and they will be stupid for a looooong time. It's not about computation power, it's the way current algorithms work are not advanced enough to create true intelligence.

[caractacus]

How can you prove that anything but yourself is self-aware?

Explain to me how self-awareness works and I'll give you an answer.

[Deleted]

How do you define "being self aware"? What truly makes us humans self aware to begin with? Our ability to think? Our ability to reason? Our ability to act with and without logic? Morals? Ethics? A soul?

If you need to think to be self aware, but have to be self aware to think, we're having us a viscious Catch-22.

Cogito ergo sum. (I think, therefore I am.) - René Descartes

[Dreamless]

You got AI wrong. It's not true intelligence. It's simulation of behaviour and you program it. Machines are stupid and they will be stupid for a looooong time. It's not about computation power, it's the way current algorithms work are not advanced enough to create true intelligence.

This is incorrect. Biological-based programming has already been successfully tested, giving software sets of basic behaviours and desires and allowing for emergent behaviour rather than programmed behaviour.

This is the future of AI research.

How do you define "being self aware"? What truly makes us humans self aware to begin with? Our ability to think? Our ability to reason? Our ability to act with and without logic? Morals? Ethics? A soul?

If you need to think to be self aware, but have to be self aware to think, we're having us a viscious Catch-22.

Cogito ergo sum. (I think, therefore I am.) - René Descartes

A good question, and one which behaviourists have answered to a degree, at least in recognizing if not explaining it.

There's a famous experiment whereby a dot is drawn on the forehead of an animal and they're exposed to a mirror to see what they do. Non-self-aware animals (the vast majority) will just react to the mirror as if it's another animal. Self-aware animals will learn that it's themselves in the mirror, and will eventually explore the unfamiliar dot on their forehead.

There are several different animals which have passed this test; many primates, as well as elephants.

[Kuntantee]

This is incorrect. Biological-based programming has already been successfully tested, giving software sets of basic behaviours and desires and allowing for emergent behaviour rather than programmed behaviour.

This is the future of AI research.

I am waiting for an academical link then. Not wikipedia, not news sites.

[Dreamless]

I am waiting for an academical link then. Not wikipedia, not news sites.

Not your personal librarian. Utilize a little google-fu.

As for your wikipedia bias - you'd be correct that the information in articles is questionable due to its ability to be freely edited, however wikipedia articles also include primary source links.

Look up behavioural ai, or behaviour-based robotics and you'll find information.

And as a suggestion for next time; the snarkiness isn't liable to get people to want to fulfill your requests. Try asking.