Why Artificial Intelligence Cannot Take Your Soul

And it's not because you don't have one......

My senior year of college I took an introduction to logic course. I had to, I needed something called an “analysis point” to fill out my transcript. But it turned out to be fun. I was enthralled by the structure of the topic, the thrill of working through proofs. But then in late fall I ran a marathon, got sick and fell far behind. The complexity of the end of semester proofs boggled me. Thankfully the professor, a young assistant, offered me an extra credit option to push my grade, if not over the top, at least into an acceptable bin.

I prepared the proofs that he gave me and we met to discuss them. I’m guessing he wanted to make sure I did them myself. After some discussion he wrote the hoped for OK on the top of each proof. He stood up and went to his blackboard. Do you want to see something? He asked.

His expression was pert, with the same curl of the lip my friends back in Detroit used to get when they invited me to do small crimes. I always said no to the crimes. I am much too sensitive and flighty, oh and moral, don’t forget moral, for life on the wrong side of the line. But I said yes that day.

The professor started with a simple logical proposition. For all objects A, A is equivalent to A. Identity exists and is sustained through time. A great starting place for building reality. He then proceeded, by a series of perfectly acceptable logical operations, to prove that for all objects A, A is not equivalent to A.

So yeah, it turns out, at least from the perspective of the science of logic, we make it all up as we go along. There is no fixed unchanging logical structure of the universe with convenient hooks where we can hang our hats and go about our business. The same principle seems to apply to mathematics. In both disciplines vast complex structures exist that can birth computers, software and various engineering feats that boggle description. But there is no one unifying conceptual structure that binds all ways of being and doing together. As the great logician Kurt Godel showed, in mathematics and logic there are always things we know that we can’t prove.

The ruling principle of the universe seems to be: First, let’s define your terms.

So how does this apply to Artificial Intelligence? The current AI debate is between the replacers, researchers who actively think that AI will replace and consume us, create better people who are whirlwinds of calculation and information manipulation and never stop for cupcakes, building sand castles, or staring at flowers, and the enhancers, who say that’s pretty neat, what shall we use it for?

The replacers say yes, we can devise an algorithm that will include every aspect of a human mind. We can create a new consciousness, or whatever you want to call it. We just haven’t devised it yet. But watch us try.

The enhancers say, cool technology bro, any plans for safeguards on applications? (So far the answer is no. Ask any author who has had their life’s work thrown into the blender of AI so college students no longer have to research their research papers.)

But back to that soul thing. Let’s look at Burke’s barely hatched theory of the soul. Imagine humans have a capacity for thought and feeling to work together, to enhance each other. When they fuse somehow, someway, we get art, inspiration or just let ourself fall in love at long last.

Like all human attributes, this fusing of heart and mind is rooted in biology. Our progress through this vale of tears and fun is guided by chemicals from the brain, and from the heart, and from all our glands, luring us and reminding us and pulling us every which way. And this biology is rooted in chemistry, in electrical impulses and ultimately in quantum processes. Kind of like a computer.

That’s where the replacers leap to their great conclusion. If we just know enough, engulf enough, own enough, we can have it all! Bwaahahaha!

If you define humans as nothing but a collection of algorithmic processes that will eventually be reproduced and surpassed by some future generation of AI, then you know how the replacers think.

But what replacers will never tell you is that any project rooted in mathematical and logical processes ultimately has to work from assumptions. There is no unshifting logical and mathematical bedrock underlying their vast structures. Just assumptions, and prejudices. (And profit motives. But that is another essay.)

If you define humans as biological entities with a capacity to think and feel, and to create magic when those two abilities work together, then I’ll suggest you are in the enhancers camp.

So, yeah. First, let’s define our terms.

On a much more sophisticated level, this argument played out for decades. The amazingly accomplished superstar computer scientist Marvin Minsky of MIT (d. 2016), argued in essence that humans are machines, rabbits are machines, dandelions are machines, and someday machines will reproduce every aspect of being human. Physicist Roger Penrose, author of The Emperor’s New Mind, has long held that there is something called consciousness that cannot be reproduced or replaced. Something that takes shape in human thought processes that cannot completely be reproduced by algorithms.

Burke’s corollary to Penrose’s argument is that any definition of consciousness must at some level include the union of thought and feeling. This is what lets us be kind, hug a bunny, hold a hand. Eat that cupcake. Or, at the highest levels, perhaps someday revive the natural systems of Earth and build sustainable societies where exploitation and industrialized mass cruelty are something we read about in history files. (AI could be a great ally in that project. I’m an enhancer, not a luddite.)

Remember how people who are revived after dying almost always say the one thing that is important in this life is how you answer the question “Were you kind?” ( See my post Caregiving and the Near Death Experience.) That is an enhancer argument for understanding this crazy world.

We’ll give the replacers the last word. Kindness. Schmindness. All that human stuff will someday be encoded. My brain, your brain, all brains will be uploaded and absorbed into a super AI that will mimic, sorry, reproduce, every thought, feeling and impulse of every person over all time. This will be used to create a biological world where all this can play out in real time!

Oops. Drifted into a science fiction novel there.

So, like all the biggest problems in math and logic, whether we let AI steal our souls will come down to let’s define your terms. Which world do you want to live in? It’s your choice. You only have to keep thinking about how you think and feel. QED (The fancy math way of saying “I think I got it!) AI, cannot take your soul. Unless you give it away, or pretend you don’t have one.

Further Reading:

So much out there, but I have to give a couple shout outs.

Artificial Intelligence: A Guide for Thinking Humans By Melanie Mitchell, gave me a foothold on thinking about this bold new frontier. Mitchell leans into understanding as the special sauce that divides us from machines.

The Emperor’s New Mind by Roger Penrose makes a painstakingly worked out argument that current physics cannot fully explain the workings of consciousness.