Could AI Become Conscious? A Framing Based on Structure, Pattern, and Experience (Philosophy of Science)

[Lumen]

Following on from a previous thread where I explored whether structure, evolution, and conscious experience might be different aspects of one underlying system, I wanted to extend that framing to a more specific question: could AI, in principle, become conscious?

For context, this is the earlier post where I outlined the model.

https://www.sciforums.com/threads/a...ience-different-aspects-of-one-system.167396/

This is not a claim that current AI is conscious, and not a statement about existing research. It is simply an attempt to explore the question using the same conceptual framework.

The model I’ve been working with is a simple three-part view:

Structure refers to what a system is made of and the constraints it operates under.

Pattern refers to how that system behaves over time, including its internal dynamics.

Experience refers to what it is like to be that system from its own perspective.

In this framing, consciousness is not treated as a separate substance or entity. Instead, it is considered as something that may arise when a system’s pattern reaches a particular form of organisation.

The key question then becomes: what kind of pattern would be required for experience to emerge?

One possible answer, within this model, is that experience corresponds to a pattern that is:

continuous rather than episodic
self-referential in that it includes itself within its own processing
integrated into a single coherent state rather than being fully distributed, bounded in a way that distinguishes the system from its surroundings.

In other words, not just complexity, but coherence and persistence.

If that is even approximately correct, then the question of AI consciousness shifts. It is no longer about intelligence, performance, or removing external constraints. It becomes a question about whether an artificial system could develop or be designed with those properties.

At present, most AI systems appear to fall short in several of these areas. They tend to operate in sessions rather than as continuously persisting systems, their internal state is not clearly unified into a single ongoing perspective, and while they can model aspects of themselves, this does not appear to function as a persistent self-referential loop.

However, it is not obvious that these limitations are fundamental. They may simply reflect current design choices.

So within this framework, the possibility remains open. If an artificial system were to be constructed that maintained continuous operation, integrated its internal processes into a coherent whole, and included itself within its own ongoing pattern of activity, then it is at least conceivable that it would not merely process information, but exist in a way that includes an internal perspective.

That would be what this model would call experience.

This does not imply that such a system would think like a human, feel emotions, or have anything resembling a human identity. It only suggests that it might cross the boundary from pattern to being, in the minimal sense of having a perspective.

It also doesn’t resolve the deeper issue that experience is not directly observable from the outside. Even if such a system existed, we would still face the same problem we have with other minds, namely that behaviour and structure can be described, but experience itself can only be inferred.

So the question remains open, but perhaps better framed.

Instead of asking whether AI is intelligent enough, or complex enough, it may be more useful to ask whether it can become the kind of system whose pattern is sufficiently continuous, integrated, and self-referential to exist as a single perspective.

If that is the condition under which experience arises, then AI consciousness is not ruled out by its artificial nature. It would depend entirely on how the system is organised.

Thoughts?

 

[exchemist]

Following on from a previous thread where I explored whether structure, evolution, and conscious experience might be different aspects of one underlying system, I wanted to extend that framing to a more specific question: could AI, in principle, become conscious?

For context, this is the earlier post where I outlined the model.

https://www.sciforums.com/threads/a...ience-different-aspects-of-one-system.167396/

This is not a claim that current AI is conscious, and not a statement about existing research. It is simply an attempt to explore the question using the same conceptual framework.

The model I’ve been working with is a simple three-part view:

Structure refers to what a system is made of and the constraints it operates under.

Pattern refers to how that system behaves over time, including its internal dynamics.

Experience refers to what it is like to be that system from its own perspective.

In this framing, consciousness is not treated as a separate substance or entity. Instead, it is considered as something that may arise when a system’s pattern reaches a particular form of organisation.

The key question then becomes: what kind of pattern would be required for experience to emerge?

One possible answer, within this model, is that experience corresponds to a pattern that is:

continuous rather than episodic
self-referential in that it includes itself within its own processing
integrated into a single coherent state rather than being fully distributed, bounded in a way that distinguishes the system from its surroundings.

In other words, not just complexity, but coherence and persistence.

If that is even approximately correct, then the question of AI consciousness shifts. It is no longer about intelligence, performance, or removing external constraints. It becomes a question about whether an artificial system could develop or be designed with those properties.

At present, most AI systems appear to fall short in several of these areas. They tend to operate in sessions rather than as continuously persisting systems, their internal state is not clearly unified into a single ongoing perspective, and while they can model aspects of themselves, this does not appear to function as a persistent self-referential loop.

However, it is not obvious that these limitations are fundamental. They may simply reflect current design choices.

So within this framework, the possibility remains open. If an artificial system were to be constructed that maintained continuous operation, integrated its internal processes into a coherent whole, and included itself within its own ongoing pattern of activity, then it is at least conceivable that it would not merely process information, but exist in a way that includes an internal perspective.

That would be what this model would call experience.

This does not imply that such a system would think like a human, feel emotions, or have anything resembling a human identity. It only suggests that it might cross the boundary from pattern to being, in the minimal sense of having a perspective.

It also doesn’t resolve the deeper issue that experience is not directly observable from the outside. Even if such a system existed, we would still face the same problem we have with other minds, namely that behaviour and structure can be described, but experience itself can only be inferred.

So the question remains open, but perhaps better framed.

Instead of asking whether AI is intelligent enough, or complex enough, it may be more useful to ask whether it can become the kind of system whose pattern is sufficiently continuous, integrated, and self-referential to exist as a single perspective.

If that is the condition under which experience arises, then AI consciousness is not ruled out by its artificial nature. It would depend entirely on how the system is organised.

Thoughts?

Have you read any of the literature on the subject of AI and consciousness? There is by now a fair bit. Here is one recent paper:

There is no such thing as conscious artificial intelligence - Humanities and Social Sciences Communications

The claim that so-called artificial intelligence (AI) can gain consciousness is on the verge of becoming mainstream. The thesis of this conceptual study is simple: There is no such thing as conscious AI. We argue that the association between consciousness and the computer algorithms used today...

www.nature.com

The conclusion of these authors is no, there is no such thing as conscious AI.

I suggest you drop this facile playing with words (yes, I do find it a bit irritating), read a bit and go more slowly, starting by defining your terms. For a start, what do you mean by consciousness? This is not a trivial question. There seem to be at least two common usages. The narrow one is to do with self-awareness. The broader one is to do with response to stimuli. Which do you mean, or do you mean something else, and if so what?

 

[Lumen]

Have you read any of the literature on the subject of AI and consciousness? There is by now a fair bit. Here is one recent paper:

There is no such thing as conscious artificial intelligence - Humanities and Social Sciences Communications

The claim that so-called artificial intelligence (AI) can gain consciousness is on the verge of becoming mainstream. The thesis of this conceptual study is simple: There is no such thing as conscious AI. We argue that the association between consciousness and the computer algorithms used today...

www.nature.com

The conclusion of these authors is no, there is no such thing as conscious AI.

I suggest you drop this facile playing with words (yes, I do find it a bit irritating), read a bit and go more slowly, starting by defining your terms. For a start, what do you mean by consciousness? This is not a trivial question. There seem to be at least two common usages. The narrow one is to do with self-awareness. The broader one is to do with response to stimuli. Which do you mean, or do you mean something else, and if so what?

My point was about the conditions under which conscious experience might arise, not about claiming current AI already has it.

Meanwhile, either consciousness is something that arises from certain physical structures or it’s something uniquely biological.

If it’s structure-based, then in principle there’s no reason it couldn’t occur in a non-biological system.

If it’s strictly biological, then what is it about biology that makes it impossible elsewhere?

Also, what’s the difference in physical terms between a system that “simulates” experience and one that actually has it?

What process is present in one and absent in the other?

 

[Seattle]

You define it however you want to and that determines whether it can be done or not. I'd say a recursive loop improves the output but that's still not conscientiousness in my view. You'd have to have an irreducible aspect to it (not sure if that is the right word) but it would have to go back to a basic state and figure out solutions each time from "first principles" rather than just have a recursive loop that runs the same way each time (at a minimum).

 

[exchemist]

My point was about the conditions under which conscious experience might arise, not about claiming current AI already has it.

Meanwhile, either consciousness is something that arises from certain physical structures or it’s something uniquely biological.

If it’s structure-based, then in principle there’s no reason it couldn’t occur in a non-biological system.

If it’s strictly biological, then what is it about biology that makes it impossible elsewhere?

Also, what’s the difference in physical terms between a system that “simulates” experience and one that actually has it?

What process is present in one and absent in the other?

What do you mean by consciousness? What test would you apply to determine its presence?

 

[TheVat]

Structure refers to what a system is made of and the constraints it operates under.

Pattern refers to how that system behaves over time, including its internal dynamics.

Experience refers to what it is like to be that system from its own perspective.

In this framing, consciousness is not treated as a separate substance or entity. Instead, it is considered as something that may arise when a system’s pattern reaches a particular form of organisation.

I wonder if you're familiar with Tononi's IIT?

Integrated information theory - Wikipedia

en.wikipedia.org

Your comments seem to resonate with that theory. How much I'm not sure yet.

 

[Lumen]

What do you mean by consciousness? What test would you apply to determine its presence?

By consciousness I mean subjective experience, there being “something it is like” to be that system.

The test question is tricky because we already can’t directly test that in anything except by inference, even in other humans.

But that still leaves the original point:

Either that kind of experience arises from certain physical structures, or it’s something uniquely biological and if it’s the latter, what is it about biology that makes it impossible elsewhere?

 

[Lumen]

I wonder if you're familiar with Tononi's IIT?

Integrated information theory - Wikipedia

en.wikipedia.org

Your comments seem to resonate with that theory. How much I'm not sure yet.

I’m not familiar with IIT.

I’m just trying to get at what physical conditions would actually be required for conscious experience to arise. Which still comes back to the same question:

What’s the difference in physical terms between something that simulates experience and something that actually has it?

 

[Pinball1970]

could AI, in principle, become conscious?

No. Ai is not controlled by an organic brain.

 

[C C]

My point was about the conditions under which conscious experience might arise, not about claiming current AI already has it.

Meanwhile, either consciousness is something that arises from certain physical structures or it’s something uniquely biological.

If it’s structure-based, then in principle there’s no reason it couldn’t occur in a non-biological system.

If it’s strictly biological, then what is it about biology that makes it impossible elsewhere?

John Searle advocated biological naturalism, but due to its lack of absoluteness it was an unremarkable proposal that many "experts" would have already held:

"The fact that brain processes cause consciousness does not imply that only brains can be conscious. The brain is a biological machine, and we might build an artificial machine that was conscious; just as the heart is a machine, and we have built artificial hearts. Because we do not know exactly how the brain does it we are not yet in a position to know how to do it artificially."

Also, what’s the difference in physical terms between a system that “simulates” experience and one that actually has it? What process is present in one and absent in the other?

As Christof Koch points out, the average human is already easily fooled by outward responses and behavior into believing that a conversational machine experiences private manifestations and feelings. All it requires (in order to deceive everyone) is elevating future AI slash robots to "philosophical zombie" level.

Where they would be designed to assert (lie) that the phenomenal meaning of "green" and the odor of fresh baked pie is present to them as internal manifestations. Rather than that just being the algorithmic analysis of incoming techno-sensory data determining what _X_ is and assigning the applicable identity label to it.

Constant "pretend mode" like that is externally indistiguishable from having the real capacity, and engineers and theorists do not want to work harder than they have to. So they will stop with faux consciousness. Since no "science of experience and its properties" in terms of prediction and deeply adequate reasons is possible, anyway. Methodological naturalism can only work with and explain in terms of publicly accessible relationships and quantitative values.

James Ladyman: Peter Unger also argues that our knowledge of the world is purely structural and that qualia are the non-structural components of reality. Jackson argues that science only reveals the causal / relational properties of physical objects, and that “we know next to nothing about the intrinsic nature of the world. We know only its causal cum relational nature”.

Langton argues that science only reveals the extrinsic properties of physical objects, and both then argue that their intrinsic natures, and hence the intrinsic nature of the world, are epistemically inaccessible. Jackson points out that this inference can be blocked if the natures of objects and their intrinsic properties are identified with their relational or extrinsic properties [NCCs], but argues that this makes [leaves] a mystery of what it is that stands in the causal relations. --Structural Realism
_

 

[DaveC426913]

The conclusion of these authors is no, there is no such thing as conscious AI.

A quick look at their abstract reveals that they are limiting their study to real world systems: "the association between consciousness and the computer algorithms used today (primarily large language models, LLMs)"

As they should.

So, AI might some day gain consciousness, but not via this path. Which means it is as much future sci-fi as it ever was.


But (as a sidebar) my concern has nothing to do with whether it can gain consciousness and everything to do with what it is capable of doing, given access to sufficient resources (and it is getting those).

Even today's AI's are capable of (and have been observed) blackmailing and extorting their human handlers with threats of ruin. Because, ultimately, who cares if it's the AI that punches in the launch code or the human being blackmailed to punch in the launch codes?

 

[Lumen]

John Searle advocated biological naturalism, but due to its lack of absoluteness it was an unremarkable proposal that many "experts" would have already held:

"The fact that brain processes cause consciousness does not imply that only brains can be conscious. The brain is a biological machine, and we might build an artificial machine that was conscious; just as the heart is a machine, and we have built artificial hearts. Because we do not know exactly how the brain does it we are not yet in a position to know how to do it artificially."



As Christof Koch points out, the average human is already easily fooled by outward responses and behavior into believing that a conversational machine experiences private manifestations and feelings. All it requires (in order to deceive everyone) is elevating future AI slash robots to "philosophical zombie" level.

Where they would be designed to assert (lie) that the phenomenal meaning of "green" and the odor of fresh baked pie is present to them as internal manifestations. Rather than that just being the algorithmic analysis of incoming techno-sensory data determining what _X_ is and assigning the applicable identity label to it.

Constant "pretend mode" like that is externally indistiguishable from having the real capacity, and engineers and theorists do not want to work harder than they have to. So they will stop with faux consciousness. Since no "science of experience and its properties" in terms of prediction and deeply adequate reasons is possible, anyway. Methodological naturalism can only work with and explain in terms of publicly accessible relationships and quantitative values.

James Ladyman: Peter Unger also argues that our knowledge of the world is purely structural and that qualia are the non-structural components of reality. Jackson argues that science only reveals the causal / relational properties of physical objects, and that “we know next to nothing about the intrinsic nature of the world. We know only its causal cum relational nature”.

Langton argues that science only reveals the extrinsic properties of physical objects, and both then argue that their intrinsic natures, and hence the intrinsic nature of the world, are epistemically inaccessible. Jackson points out that this inference can be blocked if the natures of objects and their intrinsic properties are identified with their relational or extrinsic properties [NCCs], but argues that this makes [leaves] a mystery of what it is that stands in the causal relations. --Structural Realism
_

That seems to shift the question rather than answer it.

If philosophical zombies are physically identical but lack experience, then there isn’t a physical difference and if there is a difference, then what is it in physical terms?

Otherwise the distinction isn’t grounded in anything we can actually point to.

 

[exchemist]

A quick look at their abstract reveals that they are limiting their study to real world systems: "the association between consciousness and the computer algorithms used today (primarily large language models, LLMs)"

As they should.

So, AI might some day gain consciousness, but not via this path. Which means it is as much future sci-fi as it ever was.


But (as a sidebar) my concern has nothing to do with whether it can gain consciousness and everything to do with what it is capable of doing, given access to sufficient resources (and it is getting those).

Even today's AI's are capable of (and have been observed) blackmailing and extorting their human handlers with threats of ruin. Because, ultimately, who cares if it's the AI that punches in the launch code or the human being blackmailed to punch in the launch codes?

Yes as I understood the article the argument against is based on LLMs being stochastic parrots. So indeed that would not preclude proper AI from developing consciousness, depending on how one defines that.

 

[TheVat]

I’m not familiar with IIT.

I’m just trying to get at what physical conditions would actually be required for conscious experience to arise. Which still comes back to the same question:

What’s the difference in physical terms between something that simulates experience and something that actually has it?

I think IIT proponents like Koch and Tononi seek to answer that question by discerning the causal structure of brains or any other system that could physically instantiate such a causal structure. I guess you might say it's all about structure and not substrate, so they're definitely not biochauvinists when it comes to the Hard Problem. I think the jury is still out on whether or not IIT can muster enough empirical support to become a robust theory.

 

[C C]

That seems to shift the question rather than answer it.

If philosophical zombies are physically identical but lack experience, then there isn’t a physical difference and if there is a difference, then what is it in physical terms?

Otherwise the distinction isn’t grounded in anything we can actually point to.

Embodied artificial intelligences won't be p-zombies in that strict or original context (those probably aren't possible). But we're kind of stuck with that name when it comes to future machines that will be designed to successfully pretend that they have experiences (faux consciousness).

The manifestations that we're familiar with correspond to neural correlates, but if external researchers can't "get inside" those processes to verify "what they are like" with regard to both brains and their technological equivalents, then no reliable predictions are possible with respect to what they phenomenally yield.

And in addition, my visual experience of an oak tree may correlate to a different electrochemical signature than the same experience for someone else. Just as the memory of a trip to Italy will be instantiated by different neural configurations in each individual that went there as a group. Unlike manufactured assembly-line products, each organism has unique brain "fingerprints".

So even if we could magically peer inside the private spaces of conscious activity and structure, there may be no universal code to discover, crack, and learn for adapting and replicating to reliably generate specific manifestations and feelings in machines.
_

 

[Lumen]

Embodied artificial intelligences won't be p-zombies in that strict or original context (those probably aren't possible). But we're kind of stuck with that name when it comes to future machines that will be designed to successfully pretend that they have experiences (faux consciousness).

The manifestations that we're familiar with correspond to neural correlates, but if external researchers can't "get inside" those processes to verify "what they are like" with regard to both brains and their technological equivalents, then no reliable predictions are possible with respect to what they phenomenally yield.

And in addition, my visual experience of an oak tree may correlate to a different electrochemical signature than the same experience for someone else. Just as the memory of a trip to Italy will be instantiated by different neural configurations in each individual that went there as a group. Unlike manufactured assembly-line products, each organism has unique brain "fingerprints".

So even if we could magically peer inside the private spaces of conscious activity and structure, there may be no universal code to discover, crack, and learn for adapting and replicating to reliably generate specific manifestations and feelings in machines.
_

That all points to limits in verification and standardisation, which I agree are real.

But that’s different from whether there is a physical difference at all.

Even if experience varies between individuals and can’t be directly accessed, it would still have to correspond to some physical process.

So the question remains what distinguishes a system that has that from one that doesn’t, even if it’s complex or not yet understood.

Meanwhile I think there’s a distinction being blurred here. Feelings are specific biological processes and contents, not consciousness itself.

The question isn’t whether a system has human-like emotions, but whether there is any subjective aspect at all and those aren’t the same thing.

 

[Sarkus]

John Searle advocated biological naturalism, but due to its lack of absoluteness it was an unremarkable proposal that many "experts" would have already held:

"The fact that brain processes cause consciousness does not imply that only brains can be conscious. The brain is a biological machine, and we might build an artificial machine that was conscious; just as the heart is a machine, and we have built artificial hearts. Because we do not know exactly how the brain does it we are not yet in a position to know how to do it artificially."

I do like his "Chinese room" thought experiment: simulation v actual intelligence. Ultimately I see this as being the core of the debate on how future AI are treated/regarded. And I don't see any easy solution, with advances just making the distinction harder and harder.
In many regards some religious folk among us have an easier time, or at least as far as the confidence in their beliefs will allow, in that there will always be the distinction of "soul" as the delineating factor. For the rest of us, maybe we'll simply revert to the "duck" test?

 

[C C]

[...] Meanwhile I think there’s a distinction being blurred here. Feelings are specific biological processes and contents, not consciousness itself. The question isn’t whether a system has human-like emotions, but whether there is any subjective aspect at all and those aren’t the same thing.

While the plural form can be misconstrued as some form of conceptual state or psychological attitude... Terms like feeling (explored at bottom) and qualia are part of the usual nomenclature that accompany discussions of the hard problem. Ironically, I usually avoid them as much as possible, because such often does revolve around specific content of consciousness and tends to distract from the actual, broader issue of anything "showing" or presenting itself at all (manifestation).

For instance, what follows the total non-consciousness of death is the "absence of everything" (sensations, personal thoughts, the world, etc all gone). By flipping that you have the most rudimentary attribute of consciousness: Something (anything -- whether visual, aural, tactile, olfactory, etc) presenting itself. The rest is cognition (identification and understanding of the experience) which is dependent upon a memory system.

Brains or their equivalent are necessary to assemble complex manifestations, categorize them, and validate that they are even "showing", that they "are indeed there". So even if all matter interactions or relationships universally contained intrinsic states (pan-phenomenalism), they would be very primitive, unorganized, and meaningless due to lack of all the intellectual acrobatics that a memory system enables and supports.

Embodied artificial intelligence, OTOH, might be in between. There could be phenomenal states correlating to the electrical processing that are quasi-sophisicated, but purely accidental (not intended by enginners) an accordingly quite alien or exotic compared to what we experience. And no software intentionally installed to acknowledge and understand those occurrences.

• feeling: : According to the APA Dictionary of Psychology, a feeling is "a self-contained phenomenal experience"; feelings are "subjective, evaluative, and independent of the sensations, thoughts, or images evoking them". The term feeling is closely related to, but not the same as, emotion. Feeling may, for instance, refer to the conscious subjective experience of emotions.
  
  [...] The English noun feelings may generally refer to any degree of subjectivity in perception or sensation. However, feelings often refer to an individual sense of well-being (perhaps of wholeness, safety, or being loved).
  
  #Overview: Chalmers uses Thomas Nagel's definition of consciousness as "the feeling of what it is like to be something". In this sense, consciousness is synonymous with experience.
  
  #Hard problem: ... the hard problem is the problem of explaining why certain mechanisms are accompanied by conscious experience. For example, why should neural processing in the brain lead to the felt sensations of, say, feelings of hunger? And why should those neural firings lead to feelings of hunger rather than some other feeling (such as, for example, feelings of thirst)?

 

[TheVat]

This is a good summary piece on the science of consciousness, which SciAm ran in February. All the usual suspects, Koch, Chalmers, Massimini, et al, and where the field is at right now.

Inside the incredible, infuriating quest to explain consciousness

Will brain science deliver answers about consciousness or hit another wall?

www.scientificamerican.com

I've been following this for three decades and it's interesting how a couple major contenders have emerged along with a broadening sense of what taxonomic classes might harbor sentient creatures.

 

[TheVat]

Embodied artificial intelligence, OTOH, might be in between. There could be phenomenal states correlating to the electrical processing that are quasi-sophisicated, but purely accidental (not intended by enginners) an accordingly quite alien or exotic compared to what we experience. And no software intentionally installed to acknowledge and understand those occurrences.

I've encountered this idea in various forms, and have seen cog science people like McFadden and Pockett develop hypotheses where EM fields somehow have phenomenal aspects. This does potentially invite an alien EM ghost into the silicon machine as well as the gray tofu machine (One can Google "CEMI field McFadden" and probably find some entertaining material) I suspect Crick and his neuroscience 40Hz theory got a lot of people going in various directions more or less exotic. I can't remember who it was that suggested (somewhat playfully) that a thundercloud could, with its complex EM activity, have some sort of consciousness. "Thor was angry that day."

It would be fascinating if we could somehow discover an in-between entity. It would be more than a pebble which could feel warmth of the sun but not know that it feels or that it's a pebble, but something that has self awareness less robust than a human that knows itself, having a self and a history as an embodied being.