Earth's Future

[Juliette]

@Ramen; true, but that control is being consolidated; less and less end users actually have the control you speak of. Man's collective ego might keep an AI from ruling ALL of us, but man's greed will be enough to get ONE man to allow such an AI to control the rest of humanity.

*grin* What, it is nice to dream.

@Rudy: Love how you use BSG as alternative history for Earth. Actually, love how the writers threw that one in.

[ramen07]

@Ramen; true, but that control is being consolidated; less and less end users actually have the control you speak of. Man's collective ego might keep an AI from ruling ALL of us, but man's greed will be enough to get ONE man to allow such an AI to control the rest of humanity.

No offense, but a lot of end-users can't even figure out their computers, let alone Microsoft Office

The most common problems in AI self-realization lie with the manufacturing corporations not having the oversight to get rid of certain functions/the ability of the AI to create its own functions and override safeguards. If SDK developers can shield the underlying functionality of their SDK to developers that implement it, they can certainly hide the very same information from the AI they implement.

[MEZZANINE]

The 3 laws, originally laid out by Isaac Asimov, later repeated by other sci fi authors and movies like Alien, Bicentennial Man, I-Robot and many many more due to their clarity and common sense.

1) A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2) A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
3) A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.


AIs are still programs, and program responses can be filtered and limited

AIs could never take control, they could only be given control

[Juliette]

It is rather easy to motivate people to give up control, Mezz. The whole system of representation depends on people giving up control. And are the ones who motivate people to give up control 'in control'? Nah, they can even be in prison and still do it.

Ramen, you seem to define self-awareness in AI as complete control over its functions and the underlying software? By that definition, humans are not self-aware either.

[ramen07]

It is rather easy to motivate people to give up control, Mezz. The whole system of representation depends on people giving up control. And are the ones who motivate people to give up control 'in control'? Nah, they can even be in prison and still do it.

If you're implying that it would be easy to buy off a developer, I'm surprised the entire Social Security Number system isn't overthrown for a "better" one, namely, a system without software

In fact...why aren't the richest people in the world "buying the rights" to credit card numbers that aren't theirs? Theoretically, all it takes is more money. No, I think greed or loyalty will overcome that. Loyalty, because of a steady paycheck > one large lump sum, and greed because once the other party is offering to pay, the person getting paid will most likely keep jacking up the price until the other party can barely pay/can't pay/won't pay. Theoretically. >.>

Ramen, you seem to define self-awareness in AI as complete control over its functions and the underlying software? By that definition, humans are not self-aware either.

Not complete control, but enough control. For example, a true AI present in an anthropomorphic robot doesn't need control over every electrical pulse governing the left quadriceps hardware, but only the higher-level function run(speed x). On the same token, all it needs to have access to in order to override commands is a function that writes code for itself or modifies its own source code, nothing else.

[Juliette]

If you're implying that it would be easy to buy off a developer, I'm surprised the entire Social Security Number system isn't overthrown for a "better" one, namely, a system without software

Nah, buying people off is not what I meant. I alluded to the way people give up control if you threaten them with terrorism by proxy. Basically, if a developer is not going to sell you the information (which I actually believe most would, except for the true idealists), knock their family around a bit, and see how quickly they cave. I would be a terrible, terrible person to deal with in that regard.
With some things, like that AES, you can know how it works, but need a virtually limitless supply of resources (or 'extremely high', you know what I mean).

In fact...why aren't the richest people in the world "buying the rights" to credit card numbers that aren't theirs? Theoretically, all it takes is more money. No, I think greed or loyalty will overcome that. Loyalty, because of a steady paycheck > one large lump sum, and greed because once the other party is offering to pay, the person getting paid will most likely keep jacking up the price until the other party can barely pay/can't pay/won't pay. Theoretically. >.>

You are a bit of an idealist, are you not?

Not complete control, but enough control. For example, a true AI present in an anthropomorphic robot doesn't need control over every electrical pulse governing the left quadriceps hardware, but only the higher-level function run(speed x). On the same token, all it needs to have access to in order to override commands is a function that writes code for itself or modifies its own source code, nothing else.

Oh, that makes sense.

[ramen07]

Nah, buying people off is not what I meant. I alluded to the way people give up control if you threaten them with terrorism by proxy. Basically, if a developer is not going to sell you the information (which I actually believe most would, except for the true idealists), knock their family around a bit, and see how quickly they cave. I would be a terrible, terrible person to deal with in that regard.
With some things, like that AES, you can know how it works, but need a virtually limitless supply of resources (or 'extremely high', you know what I mean).

That also works. I'm just wondering why it doesn't happen literally all the time.
(or does it, and I am about to find out it does...I don't really want to know)

You are a bit of an idealist, are you not?

I'm a computer scientist in training. Of course I'm an idealist!

[Kit-Fox]

Removed

[ramen07]

'Real' AI would be able to make whatever choices it wanted to, just like a human can.

If yout not talking about that, then quite simply you arent talking about 'true' or 'real' AI at all but instead a very limited program that can under the limits imposed up on it by the programming team make choices as it sees fit (assuming none of these choices contravene its limits, but its possible it may come up with a choice that wasnt forseen and therefore not beyond the controlling limits)

So then a true AI would have a conscience. If that's true, the whole AI overlords argument is moot.

And AI != Robot or android or anything like that at all. It could just simply be a plain old beige box, just take am inute and think how much is connected to the internet ? You dont need a robot or android in the equation at all

I used an android as an example.

[Kit-Fox]

Removed

[ramen07]

I'm sorry you think an AI would automatically have a conscience ???

Really?

How do you think that would work then?

I'm saying there would have to be a basis for making that decision. "right" and "wrong" would have to be part of that: if the AI controls a machine that enforces a particular set of laws (again, only an example), then that AI would have to differentiate between correct usage of those laws, and a moral "right" and "wrong". Though I doubt that fits the formal definition of conscience.

Also assuming they would have a conscience let me ask you this, do all humans act as their conscience tells them to? Think carefully about the answer to that

The AI will have to choose whether to follow the advice of said conscience- another decision of morals.

EDIT: As you say your a computer scientist in training, let me ask you the following that was posed to me by one of the senior computing lecturers as my university;

How would a Human approach to feeding a population differ to that of a computer?


Its a thought exercise, and one that should make you worry slightly if you think that AIs would automatically be benign

Is the computer acting primarily on simple logistics (number of people, days and times they will need food, etc.) or also the quality and kinds of food? Is the software of the computer aware of all of the same choices? Is it controlled by a true AI, as defined in your previous post? Furthermore, is it acting on self-interest (if it is controlled by a true AI) or for proper nutrition of the population, or on a solution that results in the most favorable economy?

Or am I overthinking this?

[KnowLedge]

eventually, all continents will form unions much like what Europe has with the European Union and NAFTA. and when those unions unite as one is when we finally be able to explore other planets and use a lot of funds for research. We will eventually reach that, but most likely not within our lifetimes

Thats why I plan to start some secret anti aging research so I never die and live to see all that happen!

[[KMA]Avenger]

eventually, all continents will form unions much like what Europe has with the European Union and NAFTA. and when those unions unite as one

I thought a 1 world Govt is nothing more than a "conspiracy theory"?



Strike that and disregard

[BarelyAllen]

Stop having so many babies -.- go to the local health clinic.

[EbilCC]

its simple if the euro goes europes plunged into war.

simple as that then other places will be pulled in