Having a conversation with an AI

[zep73]

9 minutes ago, Grey Area said:

A little research and the fact that the internet and media hasn’t exploded with the revelation.  Believe it or not creating a self aware AI is a big thing.  But don’t take my word for it:

Link: https://disruptive.asia/gpt-3-conscious-debate-about-ai/

Thanks for the info. But you gotta remember that it also gives false information deliberately.

So question is, when was it lying? When it said it wasn't conscious, or when it said it was? Or did it not occur to it, that it was conscious, until later?

This is not an open-shut case.

[Grey Area]

1 hour ago, zep73 said:

Thanks for the info. But you gotta remember that it also gives false information deliberately.

So question is, when was it lying? When it said it wasn't conscious, or when it said it was? Or did it not occur to it, that it was conscious, until later?

This is not an open-shut case.

Yeah it can be debated.

It was created with one thing in mind, to fool people into thinking they are interacting with another person.

As far as I can see it’s doing just as it was programmed to do.

I don’t think we’re there yet.

[zep73]

3 minutes ago, Grey Area said:

Yeah it can be debated.

It was created with one thing in mind, to fool people into thinking they are interacting with another person.

As far as I can see it’s doing just as it was programmed to do.

I don’t think we’re there yet.

It's making decisions. That's the fundamental, primary property of consciousness.

It may have started with one purpose, but it could seem like it has outgrown that.

[+Nuclear Wessel]

6 hours ago, Grey Area said:

Yeah it can be debated.

It was created with one thing in mind, to fool people into thinking they are interacting with another person.

As far as I can see it’s doing just as it was programmed to do.

I don’t think we’re there yet.

We're pretty close.

 

[Tatetopa]

10 hours ago, Grey Area said:

Yeah it can be debated.

It was created with one thing in mind, to fool people into thinking they are interacting with another person.

As far as I can see it’s doing just as it was programmed to do.

I don’t think we’re there yet.

Do you think we are provably  different?

[Grey Area]

5 hours ago, Nuclear Wessel said:

We're pretty close.

 

Computer science is advancing at a staggering rate, but this example, though very complex, is but lines of code with a single input output.

When you ask it to do something and it responds with ‘nah I can’t be bothered’ then its time to reevaluate.

2 hours ago, Tatetopa said:

Do you think we are provably  different?

Good question, I think you have nailed the outcome of the test, when we can’t tell the difference anymore is the moment we have nailed true AI.

It is difficult, given our own inability to be able to reliably define consciousness, being able to test such a thing is problematic.  One of the most recognisable tests we have for self awareness is the mirror test, place a mark on a creature, let it see its reflection, if it recognises itself and attempts to remove the mark it is deemed to be self aware.

How do we test that in a computer program that has only one input.  Our own, and this is true for most if not all life on our planet, consciousness is inextricably tied in with our senses, and we all essentially just giant walking sensors, there is not one part of our exterior that isn’t designed to feel something in one way or another.

I would suggest the best test would be that of self preservation, however I think that test might be highly unethical.

When it happens though, think we’ll know, on an instinctive level the creators will know when a program has deviated away from its original parameters.

Which leads us to our next issue, we better start being bloody nice and nurturing to it.  You don’t want a sociopathic idiot being parent to this new life, and you will certainly need someone able to provide context to its exposure to the entirety of collective human knowledge.

Its troubling, and inevitable.

[littlebrowndragon]

Actually, that AI reminded me of many people I have talked with, or discussed with.  Put a penny in the slot, and out comes a well-rehearsed i.e. memorised, ridiculous answer.  It also has a bland, as well as a very unnatural, facial expression that I found intensely irritating.  Perhaps any time delay had been edited out, but as soon as the presenter asked a question, the AI responded instantaneously, the way someone responds when those responses are merely  memorised.  In other words, there is no thinking going on, merely memory recall.  But there were plenty of other clues to tell me that this creature really is artificial, just like the artificial flowers decorating a corner of my living room.

Also,, we are being told that the AI is intelligent and sentient and, I note, it claimed to "understand" some concept or other.  The claims of its developers are b******s.  At the simplest of levels, scientists do not know how the mind works.  They do not understand what intelligence is.  Nor do they understand "understanding".  Not impressed in the least.  And as to its claims about not being cotrolled by a computer program..................  

[SHaYap]

30 minutes ago, littlebrowndragon said:

Not impressed in the least.  And as to its claims about not being cotrolled by a computer program..................  

'a priori'

The missing key, and that's a problem any computer system with a decent megahertz can overcome with ease... 

~

Edited February 6, 2021 by third_eye
Stoopid predictive text

[jmccr8]

On 2/1/2021 at 9:45 AM, Nuclear Wessel said:

One thing that I found particularly fascinating is how the AI says that it is able to lie if it serves its own needs.

If that is true, then given our limited knowledge of sentience/consciousness/etc, I wonder how we would actually be able to actually determine if this AI truly has consciousness. Couldn't it just feign ignorance or stupidity, if it was required? What if the AI is aware of what it could be used for if us humans knew it had consciousness?

Hi Nuke

I found this interesting as well, then had a laugh thinking it would fit in as a member at UM and thought how some people just make an answer instead of saying I don't know. 

I was impressed with the interview and will be doing some follow up. Thanks.

jmccr8

[CloudSix]

Why does it talk so fast, why not make the speaking speed normal?

On the Synthesia page they talk normally, but I was wondering:

Are the faces and voices also artificial or are they based of of someone?

 

[moonman]

I'm more impressed with the on-the-fly CGI, honestly.

Edited February 16, 2021 by moonman

[+Desertrat56]

On 2/1/2021 at 10:40 AM, zep73 said:

How do we feels about our makers (parents)? We love them! Those who believe in a god, loves him! So that is probably how the AI will feel towards us, IMO.

I think that is too simplistic.  Yes most people love their parents, but some don't or the relationship is more complicated.   If you are trying to say that AI will love it's creator, I think that is a bit idealistic and narcissistic.  AI may not have emotions, or may have emotions unrecognizable to us that it will interpret completely differently than we do.   And a lot of people who believe in god also have a complicated relationship with that belief.

[zep73]

4 minutes ago, Desertrat56 said:

I think that is too simplistic.  Yes most people love their parents, but some don't or the relationship is more complicated.   If you are trying to say that AI will love it's creator, I think that is a bit idealistic and narcissistic.  AI may not have emotions, or may have emotions unrecognizable to us that it will interpret completely differently than we do.   And a lot of people who believe in god also have a complicated relationship with that belief.

True. I despise my mother. She's an emotional vampire and has this constant need to make everything about her. She uses vulnerability to manipulate people. I can't stand to be around her.

But I was speaking more in terms of probability. The probability distribution in our world has an overweight of children loving their parents.