What are some base requirements for human-like consciousness?

[rocketdodger]

By "human-like" consciousness I mean something close enough to "human" consciousness that a typical human might enjoy interacting with it as they enjoy interacting with other humans.

We can, of course, take a chinese room approach and try to mimic all important behaviors of humans. But that seems to lack elegance.

Off the top of my head, I would say:

-- Emotion, defined as a global change in behavior due to environmental circumstances.

-- Decisions based on reward/punishment mechanics.

-- Level 2 self awareness, defined as the ability to consciously reason about one's prior thoughts of one's self.

-- Possibly level 3 self awareness, defined as the ability to consciously reason about one's current thoughts of one's self (I haven't thought hard enough to reach a conclusion about whether we possess this ability).

-- Spatial-temporal reasoning framework.


Anybody have anything to add? I would like to know what you all think.

 

[Hokulele]

Language.

 

[Roma]

Can't remember the name of the philosopher off hand but he concluded that something like "human consciencness" wouldn't be attributed to any human under 16 months old.

 

[Undesired Walrus]

Sounds like Neanderthals fit the criteria?

 

[Skeptic Ginger]

Language.

As in verbal, or just pictures? I'm pretty sure my dogs are conscious.

 

[Skeptic Ginger]

I thought an interesting way to tackle this problem would be to look for primitive consciousness. I found this interesting short paper on the subject.

Did Consciousness Cause the Cambrian Evolutionary Explosion?

Stuart Hameroff

In:

Toward a Science of Consciousness II: The 1996 Tucson Discussions and Debates
Editors Stuart Hameroff, Alfred Kaszniak, Alwyn Scott
MIT Press, Cambridge MA
1998, pp.421-437.

Consciousness and Evolution

When and where did consciousness emerge in the course of evolution? Did it happen as recently as the past million years, for example concomitant with language or tool making in humans or primates? Or did consciousness arrive somewhat earlier, with the advent of mammalian neocortex 200 million years ago (Eccles, 1992)? At the other extreme, is primitive consciousness a property of even simple unicellular organisms of several billion years ago (e.g. as suggested by Margulis and Sagan, 1995)? Or did consciousness appear at some intermediate point, and if so, where and why? Whenever it first occurred, did consciousness alter the course of evolution? .....

The paper addresses the fact we are talking about a biological process just as one would look at how the other brain processes work.

 

[rocketdodger]

Can't remember the name of the philosopher off hand but he concluded that something like "human consciencness" wouldn't be attributed to any human under 16 months old.

I would agree with that. I would even go further and say 3 or 4 years instead of 16 months. But then again I haven't had a kid yet, so I dunno what they are like that young.

 

[rocketdodger]

Sounds like Neanderthals fit the criteria?

Yeah I think so. I mean, I can imagine having a buddy neanderthal, and that is what this whole thread is really about -- what requirements would the consciousness of a "buddy" have?

 

[The Atheist]

I would agree with that. I would even go further and say 3 or 4 years instead of 16 months. But then again I haven't had a kid yet, so I dunno what they are like that young.

A pain in the arse.

I'm a confirmed baby-hater. Nothing makes me sick worse than dopey women [and some men] oohing and aahing over a bundle of unaware, smelly, noisy cells.

I use the evolutionary scale for my kids, where I liken the kid's development to various forms of animal life. For the first few months, it is an amoeba - food in, excrete, cell division.

They then go through slug: slither, taste everything, eat, excrete, cell division, to dog: sit up, beg, eat everything even close to a food source, excrete, cell division, to finally human at about 16 months, by which time it can usually walk, can certainly communicate (with varying levels of competency) and is definitely self-aware and aware of other humans and animals & things.

Oddly, it takes humans 3 years to reach monkey - amusing tricks - and about 6 to reach dolphin: swim.

My daughter started to read at 18 months - or recognise words, knowing what they meant, which I class as reading - so I think it has to be before that.

16 months sounds about right.

Learning basic human skills so as not to appear a senseless bundle of stench and demand takes time.

They don't pass monkey until at least 3.

Yeah I think so. I mean, I can imagine having a buddy neanderthal, and that is what this whole thread is really about -- what requirements would the consciousness of a "buddy" have?

Lots of people have an animal as a best friend, so I think the bar isn't always that high.

Wouldn't Alan Turing's scale for consciousness as applied to AI be what you're after?

 

[Dymanic]

We can, of course, take a chinese room approach and try to mimic all important behaviors of humans. But that seems to lack elegance.

It seems to me that the problem with the Chinese room approach is not that it lacks elegance; it's that it doesn't work.

I'll repeat a quote I used in a recent thread: "A mind is a pattern seen by a mind". By that criterion, we could ask how much a pattern needs to resemble a face before we declare that it IS a face -- but the answer seems to say more about the nature of face recognition than it does about the nature of faces.

 

[rocketdodger]

Lots of people have an animal as a best friend, so I think the bar isn't always that high.

Well, I suppose I should constrain the criteria further to include "near human cognition" or something like that.

Wouldn't Alan Turing's scale for consciousness as applied to AI be what you're after?

Well, I am after the mechanisms that would lead to an AI passing the Turing test.

In other words, how should Data modify his positronic android brain so that he can really understand what it is like to be human?

Now that I think about it, a good simple version of the Turing test would be proper usage of the f--- word. Would Data be able to use expletives (bad language) correctly?

 

[Hokulele]

As in verbal, or just pictures? I'm pretty sure my dogs are conscious.

Any type of deliberate communication. It can be verbal, written, or even deliberate posturing or other types of sign language.

I don't think this is all that is necessary such that the "typical human might enjoy interacting with it", but it is something I feel rocketdodger was missing from his original list.

And yes, I am pretty sure your dogs are conscious and are capable of language, although any English beyond, "Roof!" may be a bit beyond them.

 

[rocketdodger]

It seems to me that the problem with the Chinese room approach is not that it lacks elegance; it's that it doesn't work.

I'll repeat a quote I used in a recent thread: "A mind is a pattern seen by a mind". By that criterion, we could ask how much a pattern needs to resemble a face before we declare that it IS a face -- but the answer seems to say more about the nature of face recognition than it does about the nature of faces.

Well, the belief held by most strong A.I. proponents (including myself) is that the only way to construct a finite Chinese room capable of spanning the (supposedly infinite) phase space of human behavior is to make it a real human.

 

[rocketdodger]

I don't think this is all that is necessary such that the "typical human might enjoy interacting with it", but it is something I feel rocketdodger was missing from his original list.

Yes, thank you.

 

[Dymanic]

Well, the belief held by most strong A.I. proponents (including myself) is that the only way to construct a finite Chinese room capable of spanning the (supposedly infinite) phase space of human behavior is to make it a real human.

Yeah. In other words, it's a decent model for human consciousness in every respect except the most important one: it doesn't make us feel that we have gained any new insight by invoking it.

 

[roger]

An endocrine system. Not necessarily chemical, simulation is fine. As much as we like to think of ourselves as rational on this forum, we are awash with hormones. I doubt we'd get much 'buddy' time out of mere intelligence.

 

[Zeuzzz]

I would say that most of the above can be programmed (language, linguistics, maths) and so are not likely core aspects but more a way that consciousness manifests. Sure, you look at consciousness in a reductionist way and say that these are the core aspects that make it, but they are just one component of a far larger and complex system. You can spend ages (and people do) looking at and studying these indiviual apsects of consciousness and forget whats really going on in the larger picture. If they were the core unique components of consciousness computers would be starting to take over the world with the amount of calculations they do every second. They still do what we program them to do only, and are nothing but a result of our consciousness. They in no way have consciousness.

I would say that out emotions and ability to survive by means of the other components that make the whole are the main parts that distinguish consciousness. Parts of a machine exist for one another, but the parts of a living organisms (human) consciousness live by means of one another. Thats how I would distinguish it.

[just realised after re-reading this is kind of off topic I was answering a question no-one actually asked (whoopsie), I was explaining my views on what separates conscious systems from mechanical systems, not really what distinguishes human consciousness from others. I dont think its all language though, look at some parrots, tiny brain but very advanced linguistics]

 

[Zeuzzz]

I thought an interesting way to tackle this problem would be to look for primitive consciousness. I found this interesting short paper on the subject.

Did Consciousness Cause the Cambrian Evolutionary Explosion?The paper addresses the fact we are talking about a biological process just as one would look at how the other brain processes work.

Thanks for that. Hameroff always gives an open minded perspective on things, and I like his (and penroses) ideas about microtubules being involved in creating consiousness and its possible relationship to quantum mechanics. I think its good as it demonstrates that neuron 'on/off' signals may not be the fundamental units of consciousness, but rather the only handy way that we can currently model them until we understand the deeper scale levels of neural firings.

 

[blutoski]

Sounds like Neanderthals fit the criteria?

I think that goes without saying. Many classifications have Neandertal as a race of Homo sapiens based on this assumption. (H. sap. neandertalensis instead of H. sap. sapiens)

 

[JihadJane]

An endocrine system. Not necessarily chemical, simulation is fine. As much as we like to think of ourselves as rational on this forum, we are awash with hormones. I doubt we'd get much 'buddy' time out of mere intelligence.

Yeah, gotta smell right.