Jump to content
Objectivism Online Forum

Rights of Artificial Intelligence

Rate this topic


VECT

Recommended Posts

'If you infer X from Y then you will be wrong, if Y is ultimately false'.  Um. . .  Yes?  The only sense I can make of that statement is that it's meant to imply "you will simply be wrong," along the same lines as "computers will appear to be conscious soon, but won't really be"- except that I can't understand that, either.

 

Whatever it is that connects those dots for you guys, it's not in my vocabulary. 

 

I haven't followed that closely yet the conversation between you and dream_weaver, but I can elaborate a bit more about my own reasons for stating why the Turing Test is not sufficient.

 

First of all, it is possible to develop a test that is sufficient to validate based purely on output whether or not an AI is truly a replication of human conciousness rather than an imitation.

 

Such a test would need to evaluate the AI's potential to learn/understand new knowledge and it's ability for applying newly learned knowledge in creative endeavours such as art, music, solution design..etc.

 

The Turing Test, while still an impossible hurdle for modern day AI, is far too shallow of a test in the large scheme of things. Reactively answering questions in an attempt to fool humans in a short time period can be accomplished with a large enough strategically organized data base and a set of efficient query algorithm.

Edited by VECT
Link to comment
Share on other sites

I was actually asked recently what would be my opinion on Artificial Intelligence and robots with conciousness and stuff like that. My answer simply was that it is possible as long as that robot or computer were volitional beings capable of existing without the need of a user or anything.

 

In short, when we get computers that are ends in themselves, we are going to have A.I.

Link to comment
Share on other sites

@Devil's Advocate:

 

Apologies, completely missed your post, here are my late replies:

 

Our agreement highly favors volitional actions and self-sufficiency for recognizing a right to life.  But there remains some fleshing out to determine what behavioral distinction, if any, programming has over instinct in terms of positing volitional action.

 

If the AI have volition, that means it will have the ability to have the final say on it's actions despite the influence of any programmings. Example would be if we got a robot programmed to assassinate certain individual, that robot refuses to do so by an act of its own will.

 

Programmings present compulsions on a volitional AI just as natural urges influence a human being (hunger, thirst, sleepiness..etc.). The magnitude and specificness of programming compulsions can of course be far greater than compared to natural urges. But if an AI truly have volition, that means it can fight back against any programmings (whether or not it have a strong enough will to win is another story).

 

I wonder...  Do pre-programmed emotional responses activiated by the selection of pre-programmed values demonstrate the moral choices of your AI, or its programmer??

 

If the AI relies completely on pre-programmed values to activate programmed emotions, then of course that just demonstrate only the moral choices of the programmer.

 

But If the AI is actually volitional, and on top of that, have an additional module for learning (such as a rational faculty), then it can form its own values to replace any pre-programmed values by an act of will. If it succeeds in doing so, then any choice it makes and any programmed emotions triggered, will be a result of its own chosen moral values. When that's the case, the AI is no longer just a puppet to its maker.

Edited by VECT
Link to comment
Share on other sites

The mention of party trick falls short of what may well be a philosophic trick. Turing allots 5 minutes, if wiki is accurate on this aspect.

It's not. You can have all the time you want, talk about anything you want. 

 

Proof is a step by step process of establishing the relationship of all the available evidence leading to a conclusion. If the Turing test is the only test available for establishing intelligent consciousness for both human and computers alike, then the conclusion follows accordingly. This does make me wonder how humans were determined to be either intelligent or conscious prior to the development of the Turing test in 1950.

The same way the Turing test does: through interaction. The "Turing test" isn't the test, it's a description of the conditions under which an unbiased experiment should be performed to administer any test you wish to administer. A better name for it would be the Turing method.

 

The only rules are that you are restricted to interacting with the subjects of your tests through a screen and keyboard (at least until technology can render speech well enough), and the humans involved must be total strangers (the person doing the testing can be a skeptic, but the other participants must be neutral, preferably not even know what the subject of the testing is).

 

The reason for the first rule is because having a human face or voice isn't a requirement for intelligence. We recognize intelligence through verbal interaction, not sight, sound, or splitting open people's skulls and looking inside. The double blind control eliminates any bias for or against the computer, and assures that it is judged by the same standard the human subject is. It's a standard scientific method aimed at eliminating bias.

Edited by Nicky
Link to comment
Share on other sites

It's not. You can have all the time you want, talk about anything you want. 

 

Then the question boils down to: Does the inability to distinguish exchange between the output of a computer executing a program and the output of a human being mean there is no difference? To that, the reply is clear: There is no difference that I can discern.

 

(Ok, Nicky, who programmed you, and what platform are you running on? :worry:)

((On another note: What philosophy would such programming develop, or would it implicitly adhere to the one it was programmed with? Fodder for another thread.))

Link to comment
Share on other sites

... here are my late replies:

 

... If the AI is actually volitional, and on top of that, have an additional module for learning (such as a rational faculty), then it can form its own values to replace any pre-programmed values by an act of will. If it succeeds in doing so, then any choice it makes and any programmed emotions triggered, will be a result of its own chosen moral values. When that's the case, the AI is no longer just a puppet to its maker.

 

That's a big if, but yes, having the ability to replace any pre-programmed values by an act of will would effectively snip any strings to the creator of its program.  So many intersting aspects of this to consider, but in terms of our agreement, we remain copacetic on this issue.

 

As suggested by Harrison Danneskjold, I do harbor the premise that such an AI would probably need to incorporate some wetware to complete the transition from a lifeless computer to a living entity, some human DNA for example, but the primary test would be to achieve volition.

 

I'm also enjoying the conversation between Harrison, dream-weaver & Nicky.

Link to comment
Share on other sites

I was actually asked recently what would be my opinion on Artificial Intelligence and robots with conciousness and stuff like that. My answer simply was that it is possible as long as that robot or computer were volitional beings capable of existing without the need of a user or anything.

 

In short, when we get computers that are ends in themselves, we are going to have A.I.

 

Agreed, and welcome to this forum :thumbsup:

Link to comment
Share on other sites

@Devil's Advocate

 

It could be the case that volition is a characteristic only reproducible in carbon based organic compounds.

 

Though I have a feeling that might not be the case; the trick of it has something to do with decentralization. Nanobots used currently as a substitute for organic cells might be the key here in reproducing volition outside of organic medium.

 

But all these are just personal guesses.

Link to comment
Share on other sites

Then the question boils down to: Does the inability to distinguish exchange between the output of a computer executing a program and the output of a human being mean there is no difference? To that, the reply is clear: There is no difference that I can discern.

Your concepts are based on things you discerned, not on things you didn't discern. You're not calling the people around you intelligent based on anything you failed to discern as of yet. 

 

If you're rational, you'll call an AI the same names you call humans by, for as long as you're presented with the same discernable facts about it.

Link to comment
Share on other sites

Your concepts are based on things you discerned, not on things you didn't discern. You're not calling the people around you intelligent based on anything you failed to discern as of yet. 

 

If you're rational, you'll call an AI the same names you call humans by, for as long as you're presented with the same discernable facts about it.

Being rational, I consider all the responses received on the merit of the content of the responses. For simplicity, I can categorize the content as true, false, or arbitrary. "Intelligent" would be based on the fact that I am able to categorize the content of the response accordingly.

 

The immediate question that comes to mind, why, (other than the fact that it is part of the topic thus adding to the complexity of the analysis) would AI even come up? Presumably, it has not.

I think I am in agreement with your point here, that is, if my response does not raise an objection.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...