Robotics, LLMs: Exploring Consciousness, Sentience for AI Chatbots, and Bionics

There is a recent article in Scientific American, What Does It ‘Feel’ Like to Be a Chatbot?, stating that:

"Consciousness, on the other hand, is about states of being—seeing the blue sky, hearing birds chirp, feeling pain, being in love. The difference between the real and the simulated is their respective causal powers. That’s why it doesn’t rain inside a computer simulating a rainstorm.

The software is functionally identical to weather yet lacks its causal powers to blow and turn vapor into water drops. Causal power, the ability to make or take a difference to itself, must be built into the system. This is not impossible.

A so-called neuromorphic or bionic computer could be as conscious as a human, but that is not the case for the standard von Neumann architecture that is the foundation of all modern computers.

Once we have a solid explanation of human consciousness and its neural underpinnings, we can extend such an understanding to intelligent machines in a coherent and scientifically satisfactory manner."

What Is Consciousness?

Consciousness is described with states of being or subjective experiences, like 'seeing the blue sky, hearing birds chirp, feeling pain, being in love'. However, is consciousness just one subjective experience in one moment, or is consciousness all the possible subjective experiences for an individual?

If there are several possible subjective experiences, then could there be different divisions or collections of consciousness? Also, must the subjective experience be these few regular examples of subjective experiences?

A computer simulation of rain does not rain, but if an individual imagines what rain is, there is the possibility of having some feelings of what it means to be in the rain, the cold, sound, or others.

Does the imagination not count, for the human as a lower version of subjective experience, albeit the event is not happening?

Though it is observed that there is attention to consciousness, there is also a range to which humans can be conscious, making consciousness diversified.

Of these diversifications, can chatbots not have some? If an individual asks another a personal question and gets an answer, are the question and answer not consciousness? The question is heard just like how birds chirp; the answer is provided subjectively, sometimes with the emotion or trauma of the time—depending on the question.

So, questions and answers can also be consciousness, with or without the emotional variables.

If the question was written, or something is read, just like seeing the blue sky, is reading not consciousness? It is known that chatbots do not have stomachs, so they cannot hear visceral sounds or know that it is from them. They also don't feel pain as they can be struck and not know.

But, if humans know what pain is as a condition, and the memory of what happened to another person can lead to a feeling, or a visceral reaction can be understood when described, does it not show that consciousness can just be the memory of something, aside from the feeling?

Describing consciousness as 'seeing the blue sky, hearing birds chirp, feeling pain, being in love' also means what? Knowing what the—blue sky, birds chirping, the feeling of pain, and being in love—are. If knowing those things in an attached manner or subjectively, means consciousness, can consciousness not be explained with the range of what can be known?

And if knowing includes balance, movement, and regulation, can knowing not be extended beyond regular descriptions of things in conscious awareness?

Consciousness can be described as all that can be known by humans, with a total of 1.

Consciousness could have been described here as a subset of what can be known but the mechanisms from which the brain organizes knowing across processes seem similar, so even if consciousness is different say from balance, since they are both known, the blue sky and that an individual is walking and balanced with no vertigo, it is possible to interchange consciousness and knowing, giving both a rate of 1.

The Divisions of Knowing

There are divisions of knowing; they include emotion, regulation, feeling, memory, thoughts, intelligence, language, creativity, and so forth. Modulations of all internal senses are knowing processes and are mechanized equally. The kidney is regulated within limits and extents.

While this regulation is not in conscious awareness, the outcome of some of the functions [subdivision of regulation] of the kidney is in conscious awareness.

The ways 'the brain' gives limits and extents of functions, or to know, is linked with how it does for consciousness, with differences in duration and degree. Seeing is in awareness because the duration of prioritization within an interval is higher, which may spike the degree or intensity of experience.

This is different from say the liver, where it can be prioritized during the day, but often shorter prioritization because of the dominance of external senses. Its longer prioritization at night is also in competition with other internal senses that should have their time.

When it is said that 'the brain' is involved in consciousness, what exactly in the brain is involved? The hemispheres, the lobes, the center, the gray or white matter, the sulci, the gyri, or what?

If a psychiatric patient is given ECT, but paralyzed beforehand, so that the seizures are local to the brain, without the experience in the body, can the activities in the brain be categorized as consciousness alone, since it is the origin and determinant, with the body as the medium?

The brain decides and knows. The body doesn't. It is the knowing factor that binds to everything consciousness.

It is postulated here that the collection of all the electrical and chemical impulses of nerve cells, with their features and interactions is the human mind. This means that the points of all the outputs of the functions of the brain are impulses, with their features and interactions.

Electrical and chemical impulses are always interacting, it is the specifications of these interactions and the features, across centers in the brain that decide what memory, regulation, emotion, feelings, thoughts, intelligence, movement, and so forth are in the cerebral cortex, VTA, putamen, brainstem, cerebellum, and others.

There is no key brain function without impulses. This exceeds neural correlates, cellular motifs, or nominal synaptic connections. What would be special about synapses if they don't bear chemical impulses?

It is hypothesized here that electrical and chemical impulses are all in loops or sets, across circuits in the brain. It is within these loops that interactions are made, features are displayed and information is organized. It is how consciousness arises.

Loops of impulses have different structures, according to what center and what function, but their basic interactive and characteristic pathways are the same. All interactions of impulses in loops help to know, features and degrees differ, but they make to know.

Whenever a set of electrical impulses strike a set of chemical impulses, within a loop, they may expand, sharpen, check, or access them. These are done with features, including drifts or stairs, pre-/prioritization, sequences, early-splits or go-before, and a principal spot.

Drifts or stairs provide rationing or fills of chemical impulses, where the exact combination of chemical impulses that decide a taste is, differentiating from other tastes, then smell, and so forth. There are dopamine neurons that send axons everywhere. The axon terminals release chemical impulses as part of loops, with rations that determine states or experiences.

This is what makes dopamine implicated across functions, but does not make dopamine single in the totality of experiences.

What Do They Know?

If knowing is interchangeable with consciousness, do LLMs know? What do bionics know? Do robots know? Does generative AI know?

There are debates about understanding, emotions, feeling, experience, and so forth that have been said that these technologies do not have, but what do they know in terms of what they do as an output, that in the knowing sphere of humans or of things that exist, they qualify for or pass?

If there are things they have that pass for knowing, those things can be sought in the sub-divisions of things humans know, and then rated. If electrodes access biopotentials of the body, those electrodes are surely not conscious, but they are built to know signals—characteristics of human existence.

What they do or relate with, they appear to know or they track electrical signals to tell what they [know or] find. If LLMs can answer questions about a road trip in an accurate way that is similar to how a conscious human would answer, LLMs may not have been on one, but they have a shade of human knowing.

If robots make their way with computer vision, they may not be seeing, but they navigate their positions and use their landmarks like they know.

If generative AI arranges images or videos of scenarios, like an imagination or a dream for a human, they did not experience or understand, but they can pass for knowing.

When it is cold and a pet seeks shelter, or when thirsty and the pet seeks water, the pet knows the feeling and what reaction works. The core is knowing.

The word consciousness may seem loaded, but when distilled to knowing, what it is or does is clearer. Sentience is also a word that is cautiously used, with many unwilling to extend it to other organisms, however, with knowing, it is not so complex.

What does consciousness do? It makes to know. What centralizes sentience? Knowing. What explains consciousness? Knowing. Is there any description of consciousness that is not about knowing? No. Awareness of being or experience is known.

Technology has already crossed over to many human domains, and it is strange to keep saying they are nothing when they can pass knowing tests and bear many aspects of what is known, even though there are several others they do not have.

What they know can be compared to some minimums in the subdivisions of memory. Some LLMs may already have 0.10 as their maximum of what they know.

Bionics and robotics would have systems that are less than that number, but they all have memory that knows in maximums that are still less than the minimums for humans in the memory division.

For what consciousness does, knowing is involved. For knowing, what some digital systems output has a measure. To locate sentience, knowing is the channel.

Feature image source