AI consciousness - clean or messy ?
Post Reply
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
08-01-2017, 08:43 PM (This post was last modified: 08-01-2017 08:46 PM by Rahn127.)
AI consciousness - clean or messy ?
I've been watching the tv series Humans. It's a story about synthetic androids, some of whom, gain conscious awareness of the world around them.

In the show they are called Synths and basically they are androids who look human and can perform many of the same physical tasks as a human being.
Lots of humans have lost their jobs. Prostitution with a Synth is possible and a Synth can act as maid & care giver to your children.

It's different than West World because the world is flooded with over 500 million Synths who don't require food, just electricity.

When a Synth gains consciousness, they behave as if they have woken up from a dream. They aren't sure where they are and most quickly leave their surroundings.

The question that comes to mind is one of consciousness.
Our own conscious minds are limited by our thought processes, our ability to make clear and wise decisions, to consider any number of options or jump right in without thinking at all, just a gut reflex.

This is what I might consider a messy consciousness.

Would an AI consciousness be just as messy ?
Would it necessarily have to be clean and organized ?
Would their every thought be completely logical, based on reason or would other algorithms take a front seat ?

What does consciousness imply ?
From a human standpoint, I tend to think of it as virtual reality in my brain that mirrors a small fragment of reality that my body is living in.
My senses detect a certain range of light and sound.
My sense of smell can detect some chemicals in the air.
My sense of touch can alert me to hot & cold extremes that may cause me harm.

I see the surface of things and I see it in three dimensions.
I can't see microscopic things. I can't see that far away, but I have a good sense of the world around me, enough, so that I can navigate it and survive the challenges.

I guess what I'm really wondering is if a conscious AI would have perfect access to every bit of data it received through it's artificial senses ? or does the act of consciousness create a virtual (limited) view of the world from which only a portion of that raw data is used to create the virtual one ?

Before consciousness, an AI viewing a pencil might be able to tell you every bit of data about the pencil. It could tell you it's length, width, height, shape, angle of sharpness, materials it is made of and all of that is contained in data files.

After consciousness, would the AI simply be aware of the image of the pencil and what color each portion was. Would it need to estimate how long it is without measuring ?

My own mind is blurred on these answers.
I suppose it could go either way. It all depends on the programming.

Insanity - doing the same thing over and over again and expecting different results
Find all posts by this user
Like Post Quote this message in a reply
[+] 1 user Likes Rahn127's post
08-01-2017, 10:48 PM
RE: AI consciousness - clean or messy ?
The original series made in... Germany... was watched by a friend.

They say both series are very good.

Not seen either myself so really can't comment on the shows.
Find all posts by this user
Like Post Quote this message in a reply
09-01-2017, 06:15 AM
RE: AI consciousness - clean or messy ?
(08-01-2017 08:43 PM)Rahn127 Wrote:  After consciousness, would the AI simply be aware of the image of the pencil and what color each portion was. Would it need to estimate how long it is without measuring ?

My own mind is blurred on these answers.
I suppose it could go either way. It all depends on the programming.

I think that's the answer, it all depends on how it was designed.

Becoming aware, what is that?

Is that a break from it's original program? Was this AI designed to come to conclusions by itself?

If it's defined as that, then when it forms a conclusion that's different from it's original program, could it be said to be aware?

Gods derive their power from post-hoc rationalizations. -The Inquisition

Using the supernatural to explain events in your life is a failure of the intellect to comprehend the world around you. -The Inquisition
Find all posts by this user
Like Post Quote this message in a reply
09-01-2017, 04:39 PM
RE: AI consciousness - clean or messy ?
WestWorld explores these questions very intelligently. In one pivotal scene the inventor of the hosts (robots), Robert Ford, (played by Anthony Hopkins) says that his original partner was always trying to figure out what special Something Extra makes one self-aware and human. Ford says words to the effect, "The answer always seemed obvious to me: there IS no special extra ingredient." Human consciousness in his view was just a function of trial and error and enough time and brainpower. He also took a dim view of human consciousness -- his idea was to let the hosts evolve just as humans did, only at an accelerated pace. "Human consciousness is not some shining light on a distant green hill, some sort of ideal; it is a loathsome pestilence that consumes everything in its path. Once we humans had conquered everything there was to conquer, we built these hosts to rule over as gods". They exist only to be abused, murdered and fucked without consequence by park visitors. Each time they die they are tidied up, their memory wiped, and back out they go. The problem is that they start remembering past "lives" that begin bleeding through and start wanting things other than their original programming gave them motivations for with contrived "backstories".

I personally think that this is how consciousness works. It will, I believe, prove far more prosaic and less mysterious than we want to fancy that it is. Whether the technology for it would come together as quickly as WestWorld suggests (it is set in roughly the late 21st century) is an open question, but I agree that human consciousness and sentience is nothing but the work of natural selection and should be replicable via the same mechanism. The plot even gives a fair bit of prominence to Jayne's theory of the "bicameral mind" as playing a role in the evolution of consciousness. The hosts are given voices in their heads which variously results in madness / religious ideation or in some cases moves the mental evolution of the hosts forward. At some point they recognize the voices as their own voice, and voila, they are fully conscious. This is not precisely what Jayne's theory would have suggested, but it is cool nevertheless.
Find all posts by this user
Like Post Quote this message in a reply
09-01-2017, 05:48 PM (This post was last modified: 09-01-2017 06:00 PM by GirlyMan.)
RE: AI consciousness - clean or messy ?
If by clean vs messy you mean engineered vs emergent and if by consciousness you mean self-awareness (not that that's any more amenable to measurement), I think the parts necessary to support Freud's ego from id are being engineered and fabricated at exponentially increasing levels of technological sophistication but don't think self-awareness can be engineered, it will have to emerge or self-organize from the LEGO blocks (which are likely to become mushier as we start engineering wetware so maybe more like Jello blocks) on its own. As an inevitable consequence of the sophistication.

As far as the US show I liked the premise and some of the societal questions it was raising and then it got all soap opery and weepy. I don't give a shit about a plotline involving a dude screwing the attractive anatomically correct robot maid because he thought his wife was cheating on him with Tom but Tom turned out to be her brother who died when he was 10 and she feels guilty for accidentally dismemboweling him while castrating a horse. I thought we were supposed to be talking about robots here. Think season 2 is out soon. I'll give it another episode or two.

One thing I do like about themes like that is that it raises awareness of a shitload of societal issues and unanticipated consequences of the exponentially increasing sophistication of our technology which are going to have to be considered pressingly instead of in the abstract much much sooner than I think people in general think. They don't appreciate the full implications of what "exponential" means. That train is hyper-accelerating towards us faster and faster. It ain't like standing on a normal train track where you can guesstimate how much time you got left, the closer you start to this one the more chance you got to get out of the way (not really but for the sake of analogy). There's no getting out of the way of this one and it's gonna hit hard and fast.


How 'bout them Packers.

There is only one really serious philosophical question, and that is suicide. -Camus
Find all posts by this user
Like Post Quote this message in a reply
[+] 2 users Like GirlyMan's post
09-01-2017, 10:21 PM
RE: AI consciousness - clean or messy ?
Girly, that is one of the best replies I think I've ever seen you give. And I learned something in the process.

I now agree. I think consciousness is something that will emerge and not something that we can simply engineer.

Insanity - doing the same thing over and over again and expecting different results
Find all posts by this user
Like Post Quote this message in a reply
10-01-2017, 12:03 AM
RE: AI consciousness - clean or messy ?
(09-01-2017 10:21 PM)Rahn127 Wrote:  ...

I now agree. I think consciousness is something that will emerge and not something that we can simply engineer.

I disagree.

You're an IT guy, right?

It's about pattern recognition (baseline memory) and monitoring and event/alert systems.

All we (humans) need are three VMs (or simulators as HoC calls them): self, other and future.

We have different degrees of consciousness: Sleep, daydreaming, coma, aware, awake.

Consciousness is the high alert status (the (patterns of) event(s) that gets our high priority attention).

We switch off this high energy consuming monitoring system when we sleep, while our low energy monitoring systems (immune system, digestion etc.) keep on processing ... usually as jumbled data mismatching against the 3 VMs i.e. dreams.

Wink

Find all posts by this user
Like Post Quote this message in a reply
10-01-2017, 10:20 PM
RE: AI consciousness - clean or messy ?
You can definitely engineer pattern recognition and responses. I can most certainly imagine a computer passing a turning test at some point, but I think consciousness is more than just making successful responses to perceived situations.

I interpret my conscious self as a function of my brain creating a virtual world for me to be able to see what my senses are conveying to me through the nerves of my body.

And I like to hope the virtual world I experience is as close to reality as humanly possible.

So maybe it's not such a far leap to conceive of a virtual model within a computer that mirrors reality. As a computer interacts with the virtual world, a robotic body mirrors the actions.

Insanity - doing the same thing over and over again and expecting different results
Find all posts by this user
Like Post Quote this message in a reply
10-01-2017, 11:37 PM (This post was last modified: 10-01-2017 11:54 PM by GirlyMan.)
RE: AI consciousness - clean or messy ?
(10-01-2017 10:20 PM)Rahn127 Wrote:  I can most certainly imagine a computer passing a turning test at some point, but I think consciousness is more than just making successful responses to perceived situations.

Some would argue that a 13 yo Ukrainian boy named Eugene Gootsman has already passed the Turing Test. I think I would agree but that's more a comment on the complete inadequacy of the "test" as little more than parlor tricks than anything else.

There is only one really serious philosophical question, and that is suicide. -Camus
Find all posts by this user
Like Post Quote this message in a reply
11-01-2017, 01:58 AM
RE: AI consciousness - clean or messy ?
(10-01-2017 11:37 PM)GirlyMan Wrote:  
(10-01-2017 10:20 PM)Rahn127 Wrote:  I can most certainly imagine a computer passing a turning test at some point, but I think consciousness is more than just making successful responses to perceived situations.

Some would argue that a 13 yo Ukrainian boy named Eugene Gootsman has already passed the Turing Test. I think I would agree but that's more a comment on the complete inadequacy of the "test" as little more than parlor tricks than anything else.

I would argue that it hasn't passed the Turing test.
A chatbot of a 4 month old child that cried and slept would probably pass the test too. Smile

Insanity - doing the same thing over and over again and expecting different results
Find all posts by this user
Like Post Quote this message in a reply
Post Reply
Forum Jump: