Consciousness and AI: Do They Mix?
I want to preface this by saying that the thoughts and ideas that would eventually become this article came as a response to the following article, “Here’s How We’ll Know an AI is Conscious” by Joel Frohlich. If you haven’t read it, please do so because it’s an amazing article. Posted in their Facts So Romantic section, it truly allows you to slip into a scifi-romantic state that speaks to one’s curiosity.
Not only that, but this article, in particular, serves as a great example as to why I have a powerful obsession with consciousness.
And while I love this article and the critical question it poses, I feel like it’s missing a critical point of what consciousness is. Let’s start with this first…
What Consciousness Is NOT
Consciousness isn’t magic and doesn’t manifest from nothing. Consider the Law of Conservation or the first law of thermodynamics:
“Energy can neither be created nor destroyed — only carted from one form of energy to another.”
Therefore, consciousness could not have shown up where it wasn’t already. This goes to say that our brains — no matter how developed or undeveloped — wouldn’t be responsible for our experiencing consciousness.
So if it wasn’t our brains, that what was it?
We still refer to the process of consciousness as having to be a mental one. The need for a brain to be present is ingrained in our ideas of it, even though some organisms react to their environments in ways outside of pure instincts or involuntary actions all the time. Plants, for instance, or jellyfish, don’t have a neurological system that we would consider typically brain-like. Objectively observing their behavior, if you can recognize that they display actions linked to awareness, and you’re willing to make an exception for them, that exception can’t stop there.
There’s been a gif circulating recently that shows a white blood cell chasing bacteria. If you want to say that is purely instinctual and just the coding of that cell, couldn’t that same argument be said for the behavior of humans, plants, or animals?
Since an AI is a series of codes created by electricity, neural networks, algorithms, and other such aspects that are given to them during creation, could it possess consciousness? It would be something that it already has somehow with the mix that it is given, or it is something that it will never have — at least until we have a better understanding of consciousness ourselves to grant it purposefully.
Degrees of Consciousness
We place limits on consciousness based on our experience of it. This is understandable and an intrinsic part of human nature. But, we need to be aware of this and refrain from putting limitations on it due to our limited experience.
This is something explained in the article, even though it says that an AI couldn’t question why noises sound different even though that’s a pretty straightforward question. If the AI were programmed to learn and better understand the stimuli it is presented with, then it wouldn’t be outside the realm of possibility for an AI to manufacture such questions, similar to our own brains, which are also wired to learn.
Learning Isn’t Consciousness Either
It isn’t our drive to learn that determines our consciousness, however. We have a higher degree of consciousness compared to some other organisms, say plants, for instance. It doesn’t take away from their experiences, but it can explain why consciousness looks different in other circumstances. Consciousness will be experienced by what the organism/molecule/body possesses, which will create its own unique perception of it.
As a child, I asked the same weird perception-based questions posed in this article: is the color that is orange to me the same orange — or color at all — as it is to someone else? What if I see green, and that’s the color I think all plants are, but you see blue, and we all call it green instead. What color are your oceans?
If computers see in coding and all input coding is subject to multiple eyes, there’s an assumption that its coded #FF0000 would be universal. The same with its #0000FF. With that logic and the ability to search and learn what “cool” and “passionate” subjectively mean, why couldn’t it answer finding a color somewhere in between? Consciousness isn’t just a form of emotional awareness or information processing. Mistaking emotions with consciousness takes away from the perception of others as emotions are their own respective experience entirely. We give emotions a definition, one that AI is fed with, so they wouldn’t be coming up with the definitions independently.
Mechanics are Separate From Consciousness
This is something that bothered me in this article. When talking about blind-sighted individuals, it mentioned that healthy eyes could feed part of the brain hidden from consciousness. Nothing is hidden from consciousness. Maybe the subconscious doesn’t directly speak with what we are aware of, but even that is consciousness — look at its name. The subconscious can be changed, developed, and unlocked, which means it’s not something that can not be touched or accessed or understood in one way or another from what we would call our “conscious” brain. Calling the process of blind people seeing as things being mechanized seems to take away from this miracle of something we obviously don’t understand and downcasting it to a level where understanding isn’t desirable — like a child asking why the sky is blue and receiving the answer that it’s just the way it is.
Alternatively. If that’s the mechanized version, what is it that regular sight is if not mechanized?
Arguably, instincts and normal life-sustaining processes are all bodily mechanics. As we discussed in the previous section regarding emotions, it doesn’t take away from consciousness, and the two are separate entirely. The mechanics involved within a body probably impact the degree to which consciousness is experienced by that organism but not if it experiences it.
So — Does AI Experience Consciousness?
Personally, I think AI’s are either already conscious just with their own unique level of perception and experience of it, or they will never be conscious. They are programmed with the data that we give them. Sure, if they are programmed to learn, their actions might mimic what we see in our experience of consciousness. Still, they themselves wouldn’t maintain that level of awareness if it weren’t already given to them (which if we knew how to program consciousness, none of this would be a question anyhow).