Feasibility of Strong A.I.

[cometbah]

Topic in honour of Ada Lovelace! (Whose birthday was yesterday)

Current focus of discussion (will be periodically updated to reflect the latest posts in the thread):

Do you think that the existence of an artificial intelligence capable of performing all known human cognitive functions is possible before the year 2112? Why / why not?

[Captain Scibra]

Considering we are already developing adaptive A.I. I think there is a fair chance it is possible.

[kimham8a]

OMG then elements AI would win against humans! 

But seriously, too many things can happen in 100 years. 100 years ago we would have never believed we would be able to talk to and see people around the world without actually seeing each other. Going to the moon was fantasy.
So yes, given that such developed AI is considered possible and not just fantasy, there is a very good chance of reaching that by 2112. Maybe by then robots will have more rights and priveliges than humans because they decided that humans are too dumb to possibly take care of ourselves.

[whatifidogetcaught?]

OMG then elements AI would win against humans! 

But seriously, too many things can happen in 100 years. 100 years ago we would have never believed we would be able to talk to and see people around the world without actually seeing each other. Going to the moon was fantasy.
So yes, given that such developed AI is considered possible and not just fantasy, there is a very good chance of reaching that by 2112. Maybe by then robots will have more rights and priveliges than humans because they decided that humans are too dumb to possibly take care of ourselves.

This. Ever since the discovery of electricity, humankind has exploded with technological amazements that we quickly go "meh" with. Chances are, when a robot is able to interact while a real live human being, the world will quickly move on.

[cometbah]

No nay-sayers? =9

[Pineapple]

If 2001: A Space Odyssey is any source of empirical evidence, I think we overestimate the growth of our abilities in emulating human intelligence. The concept itself is still mystified and ambiguous; I doubt that within a hundred years we'll see anything like HAL.

[nerd1]

Once we build AIs who can program better AIs, the question wouldn't be how fast we would get sentient AIs (really fast), the question is whether we should even make sentient AI.

[kimham8a]

Once we build AIs who can program better AIs, the question wouldn't be how fast we would get sentient AIs (really fast), the question is whether we should even make sentient AI.

But by then the best ones to think about such a dilemma is AI. LOL

Once we get AI that can program better AI, they would transcend into a state of mind incomprehensibly more sophisticated than mere, 'sentience'.

[andretimpa]

We will definitely have an AI that can pass a turing test, that is, that can emulate human behaviour in a conversation. Similarly, I think we will have AIs able to imitate human behaviour with great accuracy. It is not necessary however that such an AI have a conscience. As we as human beings have so far very little understanding on how our conscience works, it's impossible to make this kind of prediction. So, I think that the main problem is that we can only build a robot as accurate as our knowledge about human beings is and this knowledge still has some flaws, that I have no idea if they can be solved in 100 years.

TL,DR
Yes, if you mean only interaction with humans
No idea, if you mean having a conscience

[Bloodshadow]

Read "The Emperor's New Mind" by Roger Penrose.

In short, the feasibility of strong AI may be more complex a problem than what we expect. It may involve mathematical and logical concepts like computability, as well as the non-deterministic nature of quantum mechanics. I used to believe in strong AI, but now I'm not sure anymore; I'd refrain from choosing a side until more about physics and neurobiology are understood.

[memimemi]

I would say that it's not only feasible; it's inevitible.

PROVIDED:

Humans can come to agreements regarding what is or is not 'intelligence,' 'mind,' or any number of other, related, concepts.

If the question is whether there will be a human-designed machine capable of displaying behaviours consistent with our definitions of the above, then yes, yes there will.  However, chances are we won't have a standard by which to judge the birth of strong AI when it happens; strong AI will be required to help sort out the puzzle of whether it is actually strong AI, or a simulcrum thereof.

Every time a machine performs better than a human in a given field (from chess to medical diagnosis), we shift the goalposts further and further away from acceptance of machine intelligence.  The day is swiftly approaching (well before 2112 [Rush reference?]) when we, as a species, will have no domain left in which we can demonstrate clear superiority.  That will be the day that we accept strong AI. 

[leandrebs]

Can't the be 2 chechboxes: don't use THIS POWER on THIS CARD for the arena decks? Like using siphon life on a ulichard on a black dragon deck, etc, etc?