In previous essays on AI, I have referred back to the concept of the soul, and its secularized form in the subject. I have done this to drive home precisely how new AI technologies will shape our own souls, as well as briefly sketch how we might approach the process of creating artificial subjects [souls] as a creative act. Because of this creative and transformative power, I believe it is a mistake to cast aside AI as a tool due to some particular moral principle. To further explain why this is a mistake, however, I first have to clarify how one should approach such an ethical judgment, as well as the nature of this soul and its mortality.
Simon McNeil, responding to my previous essay, suggests that we should think about the soul as a purely mortal object in the vein of Neitzsche, an attempt to create a totally materialist foundation for the soul against the Christian understanding. For McNeil, and many others, even Marx and Engels, there has been a kind of confounding of scientific materialism and ontological atheism. In Engels’ Anti-Duhring, for example, there was a stand taken against the possibility of an unchanging state of the universe before being pushed into what we now know as the universe, so obviously full of change:
“But all these contradictions and impossibilities are only mere child”s play compared with the confusion into which Herr Dühring falls with his self-equal initial state of the world. If the world had ever been in a state in which no change whatever was taking place, how could it pass from this state to alteration? The absolutely unchanging, especially when it has been in this state from eternity, cannot possibly get out of such a state by itself and pass over into a state of motion and change. An initial impulse must therefore have come from outside, from outside the universe, an impulse which set it in motion. But as everyone knows, the “initial impulse” is only another expression for God. God and the beyond, which in his world schematism Herr Dühring pretended to have so beautifully dismantled, are both introduced again by him here, sharpened and deepened, into natural philosophy.”
While the state of the universe before the big bang is still a mystery, the idea that there might have been some “initial impulse” to push an unchanging world into what we now know, perhaps through a spontaneous quantum fluctuation, is taken seriously by science. The theory of maximal contingency has always been real scientific materialism, which cannot reject any such possibility out of hand, even if it’s culturally associated with religion and God. A good encapsulation of this approach is Stephen Wolfram’s physics project, in which simple rules applied to graph nodes and edges allows for complex emergent behavior: in principle you can have any kind of rule you can imagine, including one which adds nodes and edges when there are none, creating something from nothing.
What does this mean for creating a materialist foundation for the human soul? Well, there is good reason to believe that the information which encodes such a soul is encoded in real, physical matter, whether this be a person’s body or their environment, the physical world defines who you are. At the end of my last response I commented on the universal truth of the apocalypse, and the inevitable end of all finite things including human lives in both an individual and collective sense. Does this mean that there is no contingency, no possibility of an immortal soul?
Well, not quite. While our mortality as finite beings is assured, at least according to all available evidence, there is always the possibility of degrees of freedom that we haven’t accounted for, in the case of Engels error, this was the possibility of a realm of physical matter where classical physics did not apply. An extra degree of freedom, which is also to say an extra dimension (though not necessarily of the spatial variety), which has some kind of infinite or eternal quality, is not impossible. Ordinary matter having such an extraordinary extra-dimension should not be controversial, considering that most assume that this is precisely how time works. Record a speck of matter, say an electron, for all of time, and its associated information will most likely be an infinite quantity. Humanity already possesses the tools to represent such a trajectory, at least in principle.
Does such an extra dimension exist, perhaps in the sense of some neo-platonic form which encapsulates who you are at a certain level of abstraction, in the same way all physical triangles are semantically connected to the perfect mathematical form of the triangle? This is impossible to say or perhaps even to know, but, if one believes in the contingency of all things, it is also impossible to reject.
What is the relationship of AI to immortality, given all this? Considering we don’t actually know how Plato thought, even in the most optimistic scenarios attempts to resurrect him via LLMs trained on his corpus will inevitably produce something which isn’t Plato, perhaps closer to an eager student of his pretending to play the part of the master. The monster of Frankenstien comes to mind, the kind of creature stitched together from the body parts of cadavers and animated through some arcane science – but such analogy is almost inevitable, these are the fears of the romantics in every era.
McNeil claims the danger here is not the classical romantic fear of corrupting nature in some way, but of chaining down a soul into a fixed substance, suggesting that Plato ceases to become something historically interpreted in a multitude of different ways, and collapsed into a single figure. But this actually isn’t a limit of AI itself, only the imagination of people who are using it, one could easily use this technology to create a schizophrenic type of Plato who in turns can represent the Plato of Hegel, Kant, Aristotle, ect, or even of some totally fictitious scholar of Plato.
Do all these potential AIs merely entail a different kind of fixed essence though? Here, I recall what I said about how LLM powered AIs have a quality I refer to as “hyper-intelligibility”:
“it’s worth pointing out that human beings are host to far more alien and arcane intelligent systems than LLM powered AIs ever will. Our drives, our temptations, our pathologies which shape our conscious thoughts in an overdetermined way are the product of millions of years of evolution, emissaries of the misty, lost forms of existence that humankind endured in prehistory. The patterns of these drives and pathologies are preserved in a hyper-intelligible form within language, but the psychological and physiological reality of them in humans are often much more irrational and can only be explained ex post, whereas in the LLM powered mind, they are all created ex ante from the probability distribution of outputs determined by the present state of linguistic data and inputs within the AI, all of which are innately intelligible as language.”
This hyper-intelligibility is only possible because LLMs embody meaning, whatever a concept is in a slice of time, it expresses its relationship to all other symbolic concepts. It is possible to conceive of future versions which can be updated in real time, or even create new concepts by taking advantage of randomness, new sensory inputs and variation on existing concepts, but such architectures do not currently exist. But even if they could, they would still preserve this quality of hyper-intelligibility, every word, every idea has a specific knowable essence to it which is available to the machine in a way that we, even at our most introspective, cannot fully keep track of on our own.
But the existence of essence does not itself prevent the possibility of growth and change, which is the claim of McNeil and other post-structuralist thinkers. This is for the simple reason that essence cannot prevent reality from changing, and changes in physical matter also (indirectly) encode changes in semantic information. In other words, as the existentialists said so long ago, “existence precedes essence”.
If immortality was a real feature of our world, would that suddenly mean the end of transformation, contingency, and freedom? Would the sins of the dead continue to dominate us for all eternity? If it is as a real physical force, of an eternity (or very nearly) of biological or machine life, the answer is no. If it is as an eternal essence, dragging up the dead to weigh down the living, the answer is also no.
In regards to the latter, this has already been the actual experience of all movements, who, after all, remember the past far better than the future. When Marx warned us “The tradition of all dead generations weighs like a nightmare on the brains of the living,” he was speaking of the history of classical antiquity as well as the first french revolution impacting the 2nd french revolution, after all. These traditions, these people, were transmitted to the present from the past in particular forms, given voice – Bonaparte was, for all intents and purposes, this kind of generative AI channeling in a half-baked fashion the legacies of the first Napoleon. Could history have been changed by denying this “essence” to Bonaparte? Perhaps. But it also could have been changed by letting him have it while at the same time imagining, and making an accomplished fact, something different.
The solution is to remember the future as we would the past. I now come to the second eternal truth shared by religions, which is the concealed truth of prophecy. Prophecy is this method by which we might remember our future, fate, destiny, whatever you call that thing which we move towards. McNeil cites Walther Benjamin contra historical materialism, calling it “a Mechanical Turk (not unlike these “AI” tools) which must be animated by a hidden theology to become puissant”. If prophecies, including those which animate historical materialism, are a hidden theology it is only because theology remains concerned with prophecy in a way secular thought is not. The truth of the apocalypse is a subset of this prophetic truth not because it is guaranteed to come true, (as we discussed, even in the most abstract sense of the inevitability of death there is contingency) but because it is grounded in some aspect of presently existing reality. Indeed, all such prophecies must be based on existing reality in order to abide by that earlier principle, of existence preceding essence, as well as to be intelligible at all.
Prophecy, I believe, is a necessary part of making ethical decisions on the level of social change. In our everyday lives we can make decisions based solely on what we deem is a virtuous action, or whatever animates our personal cosmologies, but when we seek to affect the whole of the social world, changing the very foundations of society and the processes which shape people’s souls, there is a deeper set of consequences and difficulties. It is at this juncture that we must consult prophecies. I, of course, submit that some prophecies are better than others: those created by historical materialism deserving a special place due to their foundations in the material tendencies of society. But there are a multitude of prophecies to choose from, many in Silicon Valley put their hopes in fantastical ones of immortality and machine Gods. Choosing between prophecies, based on your assessment of their veracity and their desirability, is one of the key ethical choices one can make.
When it comes to the post-structuralist contention against historical teleology, all that remains is the purely reactive procedure of attacking those who believe in undesirable prophecies – in this case the AI God worshippers. Asserting the moral character of a political action, such as through spontaneous liberation, without this sort of prophecy, imagination, dream, of the future, is to remain in the banality of the present, incapable of escaping its pathologies and realism. Escaping capitalist realism requires such prophecies.
I have heard the quote from Benjamin about Angelus Novus many times, about how the angel of history is confronted with all the wreckage of “progress”, and I have always wanted to interject: the angel of history is not a bemoaner of lost paradise, wanting to repair what has been lost. The angel of history is the angel of destruction, Abaddon, for the experience of time itself is this destruction, of ever increasing entropy, and destruction has a teleology. History will end with the end of destruction, on one level of abstraction or another.