How to create a robot that has subjective experiences, part 4

Creating the conviction that consciousness exists

From Narrow To General AI
8 min readJul 13, 2024

This is the fourth and last entry in a series on the mechanics of conscious experiences. I recommend starting with part one, part two, and part three, as this post builds on those foundations.

The self is always defined in the larger context of the world. Interactions with that world create the shape of your self-identity.

One sometimes feels that there is something to conscious experience that transcends all these specific elements: a kind of background hum, for instance, that is somehow fundamental to consciousness and that is there even when the other components are not. — Chalmers, The Conscious Mind

Before we close this series off, there is one topic that we can’t neglect: consciousness itself. The great mystery of consciousness, as described by Chalmers in The Conscious Mind, is not so much its many properties, it is why any of it exists at all. It is notable that in his book whenever Chalmers asks “why?” the question is almost always the same: why does consciousness exist instead of nothing? To give just a small selection:

why should this [physical process] be accompanied by an experience? […]

we have little idea how it arises, or why it exists at all. How could a physical system such as a brain also be an experiencer? Why should there be something it is like to be such a system? […]

Why is all this processing accompanied by an experienced inner life? […]

Why should there be conscious experience at all? […]

Though his question is meant to be somewhat rhetorical, the answer is more obvious than he might think. But to see it, you must first take a step back and look at what humans are actually doing, rather than what they claim to be doing when they judge that they are conscious.

In a certain sense, eliminativists (those who treat consciousness as an illusion) are certainly wrong. There is something there to be seen, namely the simple fact of existence — that there is something rather than nothing. Pure existence, conscious or not, can’t be brushed off as an illusion. Those who defend dualism feel that eliminativists are telling them they don’t exist at all, in any sense, physical or spiritual. Such an argument, if it were seriously taken, would be fruitless: every mind that encountered the argument would intuitively reject it.

Anytime you contemplate the topic of your own existence, and come to the conclusion that you do indeed exist, you are using the same mental machinery that you use to judge the existence of any regular object¹. These are the same functions of motivation and attention that gauge the existence of, say, pizza or a toy, and they will respond in the affirmative if you ask if either of these effectively exists. By “effective” I mean “is it usable in the context for which I searched for it?” That is, can it be accessed by someone in the context of a given need. For example, when presented with a painting of a pipe, the question “is there a pipe?” will return different answers if your goal is to have a smoke, or if you are trying to catalogue the content of an art gallery.

The same existential questions may be asked of consciousness or of the self. The thought “I am” seems to be as unavoidable a deduction as “that apple is red”. However, as with the example of the pipe above, there must be a context of motivation or usage which causes this question to be asked in the first place, and under which the confirmation makes sense. What are you trying to use this self for? What prompted you to search for the answer? What assumptions will be used to resolve ambiguities? To truly explain how any mind can discover and understand its own existence, these are the questions we must answer, since they shape the nature of the answer.

The zealotry of consciousness

You may have noticed a curious feature of discussions about consciousness: they frequently get heated or emotional. This is surprising since, unlike politics or economics, the ontological nature of human consciousness can never have any real influence in anyone’s life. On the assumption that all humans are conscious beings, no one will perceive human beings as a species differently whether or not it turns out that we are all actually “zombies” (beings with no consciousness, but who appear conscious), as long as each of us effectively look and act like we are. The phenomenological properties of consciousness — whether they are epiphenomena, or whether they supervene on physical reality, etc. — are, even by their own definitions, inconsequential to physical existence. So where does this zealous passion in defence of the sanctity of human consciousness come from?

There is one possible answer, which I’ll try to explain through a brief aside. I may perhaps be an anomaly in that I remember the first moment I realised I was conscious, including where I was, and what I was doing at the time. I also remember the feelings associated with that moment: I experienced a removal of doubt about myself, about my identity in the larger scope of things, and about my “right” to get what I want both socially and personally. Assuming I’m not unique — and I don’t think I am — the same underlying motive appears to be the driving force behind people’s desire to discover and identify their own consciousness, even if they don’t explicitly remember or realise this fact. Having read through many defences of dualism, it seems that people conceive of conscious experience as the last bastion of personal rights and freedom from the dehumanizing encroachment of mechanical psychology.

This would explain why people are so sensitive about the specifics of a topic that is otherwise of so little consequence. They feel that the wrong answers may infringe on their notions of self-determination. There is something precious about being conscious, in contrast to the creeping “alter-determination” (determination from external forces) of cognitive science and AI. Moral arguments around the unique status of consciousness are underpinned by the fear that some — or maybe all — members of our community will be designated as non-conscious, and subsequently treated as disposable machinery or objects.

Many people are afraid to see consciousness explained because they fear that if we succeed in explaining it, we will lose our moral bearings. Maybe we can imagine a conscious computer (or the consciousness of a bat) but we shouldn’t try, they think. — Dennett, Consciousness Explained

The desire for life, for self-preservation, and for material possession are the roots of this self-understanding of our own conscious existence. This is the context in which questions like “do I exist?” are asked, and the qualifications under which ambiguous cases are resolved. It is no coincidence Buddhists claim that the ego or self have their roots in temporal desires (the desire for possession, including the possession of life), and that the removal of these desires is equated with the nirvanic extinction of the self. It would be fair to presume that a person who denied their own existence was implicitly abdicating their right to obtain any benefits from this world. These are the motives that cause you to conceive of your own existence, and drive all judgments and knowledge about it.

The mind is not a free, truth-discovering agent. There are many things that exist, but that most people will not become aware of due to a lack of interest. There must always be an interest that drives you to ask questions about your own existence, to gauge if the answers are satisfactory, and to probe them when they are not. It is quite personal; not just social, but also existential. “Am I an important person?”, “do I have a right to my own separate space and property?”, “will I continue to have experiences beyond death?”, “what am I, and why do I behave the way I do?” All these motives and more trigger and direct the discovery of your existence. They are not consequences of understanding existence, but rather the roots of that understanding itself. Even contrariness — specifically, the desire not be seen as predictable — plays a role in your self-image. Who would want to admit that they are simple machines, easy to predict?

There is little motivation for a mind to accept any judgment that categorises itself as “non-existent”. If you ever entertain any doubts about your existence — due to feeling attacked, or a momentary existential crisis caused by recognition of your own ephemeral nature — you need only pose the question to your mind, and in every sense of the word — social, physical, logical, existential — the mind would return the unqualified judgment “yes”. Any other judgment would be a sign of a faulty mind.²

Our fundamental tactic of self-protection, self-control, and self-definition is not spinning webs or building dams, but telling stories, and more particularly concocting and controlling the story we tell others — and ourselves — about who we are. — Dennett, Consciousness Explained

And so the unconscious mind emphatically signals “I exist!” and is unable to direct your beliefs any other way. As new motives and questions prompt you to further develop this understanding, to learn what exactly “it” is that exists, you always do so from your first-person perspective. Your ultimate image of your self is therefore a subjective, inside-out one. As you append features to this basic foundation, such as sensory experiences, thoughts, feelings, motives, memories, etc., you always do so under the comforting recognition that the “self” exists.

Robotic consciousness

The difference between how humans and robots understand their own existence is simply this: our understanding of our own existence is underpinned by our desire to exist. Any dumb algorithm could indicate, superficially, that it exists by checking a memory register and returning a positive response, or by logging that it is running. Nevertheless, it is difficult to accept that any AI agent really understands its own existence if it were simply hard-coded to check for that fact at regular intervals (what is known as access consciousness). It must want to check that, and an affirmative answer must be important to its motives, for us to believe that it truly understands the meaning of its own existence. All meaning is motivated.

In contrast to checking a memory register, humans build elaborate narratives around their own existence — around its consistency or stochasticity, its unity or plurality, its rights and responsibilities, and so on. The many and diverse components of this understanding (social, logical, etc) are each driven by their individual motives, such as the desire for legal protection, for social recognition and participation, or for happiness and material possession. And more can be added as social or technological circumstances change, and new philosophies of mind emerge. The desire for life-affirmation, for awareness, for deep wisdom, to know what is going on, to proclaim your freedom to any who would second-guess it, are all parts of the engine or cause out of which your belief in your conscious existence arises. Any answers you find will be shaped by the particular reason you are looking for those answers in the first place. The latter are the benchmark against which answers are accepted or rejected.

Any so-called logical deductions, like “I think, therefore I am” are subsequent rationalisations attempting to prove to other people what has already been established in one’s own beliefs. These are the real roots of one’s philosophy of consciousness. Each momentary thought or belief you form about your own existence is a superficial symptom of these deeper motivated undercurrents. And such beliefs were never a “free choice”; nothing in consciousness is.

¹ Whether you arrive at your conclusions through experience or through reasoning.

² With the exception of purely speculative contemplation. But even there, the motive determines the answer: highly pedantic or mechanistic models may deny the existence of the self because doing so is useful for their underlying purposes.

--

--

From Narrow To General AI
From Narrow To General AI

Written by From Narrow To General AI

The road from Narrow AI to AGI presents both technical and philosophical challenges. This blog explores novel approaches and addresses longstanding questions.