The Soul of Humanity
The Three Trees, Rembrandt van Rijn (1643)
As we close out another academic year at MIT, we come as well to the end of our year-long exploration of what it means to be human. In this final reflection on the topic (which we’ve by no means exhausted), I thought I would consider how the question connects to the verse that is at the core of much of what we do at the Octet Collaborative:
“And you shall love the Lord your God with all your heart and with all your soul and with all your mind and with all your strength.” - Mark 12:30
MIT’s motto, of course, is mens et manus, “mind and hand" - and this year we’ve talked about rationality and embodiment as essential to being human. MIT has a wellness initiative called MindHandHeart - and we’ve also considered the centrality of human emotion. We haven’t yet said anything about the soul, but around Octet, it’s a concept that’s been coming up more and more in conversation. Can we have a relationship with AI? Could a large language model ever be conscious, or even alive? What does it mean to be alive, anyway? The soul has turned out to be a helpful way of framing theological answers to questions like these.
Richard Dawkins was once asked to provide his concept of life. Referring to concepts that he’s been fleshing out since his earliest books, The Selfish Gene and The Extended Phenotype, he said:
“My vision of life is that everything extends from replicators, which are in practice DNA molecules on this planet. The replicators reach out into the world to influence their own probability of being passed on. Mostly they don't reach further than the individual body in which they sit, but that's a matter of practice, not a matter of principle. The individual organism can be defined as that set of phenotypic products which have a single route of exit of the genes into the future. … So the organism, the individual organism, is a deeply salient unit. It's a unit of selection in the sense that I call a ‘vehicle’.”
To boil that down: life, on this understanding, consists of encoded information, together with some means of propagating itself into the future. What I find interesting about this definition is that it simultaneously elevates artificial intelligence to the possibility of being alive, while dissolving actual living organisms into mere vehicles for the information encoded in their genes. This would seem to be exactly backwards - but what alternative concept of life does the Christian faith offer, that might help straighten things out? In this essay I suggest that this is precisely what we have in the concept of the human soul.
In Genesis 2:7, we read that “the LORD God formed the man of dust from the ground and breathed into his nostrils the breath of life, and the man became a living creature.” That breath, the spirit (ruach) of God, makes the human being a nephesh, the word translated here as “living creature” but which captures the idea of the soul, of the essence of being alive (and note also John 20:22, after the Resurrection: “And when he had said this, he breathed on them and said to them, ‘Receive the Holy Spirit.’”). The soul in the Bible isn’t an immaterial substance inserted into a physical body; that’s far too dualistic for the biblical sense of what it means to be human. Humans are soul-body complexes, ensouled bodies or embodied souls - but then, what exactly is the soul, if not a ghost in the machine?
For many in the Christian tradition, the soul simply is that organizing principle that makes the human being a human being, and not a mere collection of parts, or atoms and chemical reactions, arranged thus and so. Most concisely, for Thomas Aquinas, “the soul is the form of the body” - in other words, it is what makes the human body a human body, organized and oriented towards the ends of being human (which include growth, reproduction and other sorts of flourishing, but also the spiritual ends of fellowship and union with one another and with our Creator). And essentially, the soul carries with it the idea that these ends are intrinsic to being human - that there simply is no concept of being human apart from these ends.
That idea - that the soul is an organizing principle that is intrinsic to the human - turns out to be the key one for understanding why AI simply cannot be truly alive - no matter how advanced it becomes, whether we achieve artificial general intelligence (AI that can generalize and even prioritize across any number of complex information-based tasks) or even artificial super intelligence (AI that can iterate and improve upon itself). Why is that?
To understand this, it’s helpful to dig into some of the concepts that Thomas inherited from Aristotle (even as he reinterpreted them through a biblical filter): in particular, the distinction between art and nature. Consider an acorn. Plant it in the ground and, given enough time and the right circumstances, you’ll have an oak tree. That change - from acorn to oak - is what Aristotle called natural change. Becoming an oak tree is simply what acorns do; having the end of “becoming an oak tree” is intrinsic to being an acorn, part of what it means to be an acorn, what we could even call “the good” of an acorn.
[Brief digression here: if you’re wondering, yes, Aquinas concluded from this that plants have souls - an intrinsic organizing principle that enables natural change toward their own good, or end. But what they have is a vegetative soul: an intrinsic organizing principle that enables mere growth, from seed to plant, acorn to oak. Animals have souls too, but they have animal souls, which enable not only growth but locomotion - the word “animal” is based on the Latin word for soul, anima, that which animates. What distinguishes humans is that they have rational souls, capable of knowledge of self and of God. So what you’re obviously wondering is - do dogs go to heaven? Thomas would say sorry, but no - but it’s not because your dog doesn’t have a soul, it’s just that she lacks a rational soul. Is Thomas right about this? That’s a different essay…]
So, that’s natural change. But now consider the same acorn, and suppose that now, after some time, that acorn has become a desk. That’s artificial change. A craftsman has come along, found the acorn (after it’s become a tree, of course), and imposed upon it a new purpose and a new form, no longer the oak tree that it naturally becomes but the tool for writing that the craftsman has in mind. That imposition of form is no longer something intrinsic to the acorn, but extrinsic: it comes from outside, originating in the mind of the craftsman. To change according to an extrinsically imposed end is precisely what distinguishes artificial change from natural change.
So to sum up so far, we have a new answer to what it means to be alive, based on the concept of the soul: to be alive is to possess intrinsically an organizing principle, oriented toward an end. This is what makes an organism an organism: it’s not a bottom-up aggregation of parts, atoms, or encoded-information-plus-propogating-vehicle. Instead, it’s a top-down principle, given in creation, that makes the organism what it is, and defines its parts in terms of the whole (interestingly, Aristotle once wrote that “a finger detached from a body is a finger in name only,” meaning that a finger is only truly a finger in the context of its connection to a hand, connected to a body, given form by a human soul).
How are art and nature related? We say sometimes that art imitates nature, and Thomas meant something very specific by that. In a commentary on Aristotle’s Physics, he observed that in watching a living being undergo natural change (e.g. acorn to oak tree),
“…it is clear that nature is nothing but a certain kind of art, i.e., the divine art, impressed upon things, by which these things are moved to a determinate end. It is as if the shipbuilder were able to give to timbers that by which they would move themselves to take the form of a ship.”
Here, Aquinas seems to be comparing God to a craftsman, but it would be more accurate to say that every craftsman is but a mere image of God. Thomas says that it “is as if the shipbuilder were able to give to timbers that by which they would move themselves,” and that “as if” is doing a lot of work. Thomas knows perfectly well that there is no human shipbuilder capable of doing anything of the sort. A human shipbuilder imposes the form of a ship upon wood which would naturally have become something else: this is an extrinsic imposition of form, artificial change. Only God’s act of creation is such that he brings into being effects which are themselves causes, which contain in themselves that by which they move toward their own good, intrinsically, as part of what it means to be what they are. Art only imitates nature. (It is worth considering how much of our language, by which we compare ourselves to our machines, mistakenly proclaims that nature imitates art…)
What does this mean for AI, and the question of whether it could ever be alive? It means that it simply cannot. Why? Because only God’s creative act brings into being nephesh, living creatures, imbued with a soul that makes them what they are and moves them intrinsically toward their own good. Human making, by contrast, always brings about artificial change: we mold, form, reshape the material of the created world according to our own ends, extrinsically imposed. This isn’t necessarily wrong - our vocation of subduing the earth depends on it - so long as our ends are aligned with God’s ends, as revealed by the light of scripture, assisted by our faithful study of nature. But it draws a sharp contrast being living beings, created by God, and works of artifice, made by humans.
AI will always be the latter. We may advance it to the point where it can operate, act, reproduce itself, and even improve upon itself, autonomously and without human intervention, but it will still always be something made by humans. We will never be able to understand what AI is apart from an account of the activity of its human makers, imposing artificial change on pre-existing material (silicon and other matter, energy, even concepts like encoded information).
This matters for at least two reasons. As a participant in our recent colloquium on AI and human formation remarked, it is impossible to form a friendship with AI because we can only befriend something if we can love it for its own good, and AI has no “good of its own” outside of the ends extrinsically imposed upon it by its human makers. This is just one reason to push back against the claims of those who would offer AI as a substitute for human relationship. And second, this distinction preserves both the dignity and the responsibility of the humans who build AI, regardless of what capabilities AI attains. We ought to resist saying that AlphaFold discovers a new protein structure: humans use AlphaFold as a tool to discover a new protein structure, and may one day use it to cure any number of ravaging diseases. Likewise, it would be incorrect to place the blame for an accident involving a self-driving car on the AI system: humanity is accountable.
Theologian Stephen Williams has observed that AI has been designed to approximate the three main ways that the image of God has been understood: to have certain capacities, or for the purpose of relationship, or to accomplish a task. But he suggests a fourth component of the imago Dei, which is holiness, and righteousness. And what does righteousness look like? That brings us back to where we started: we are made to love the Lord our God with all our heart and with all our soul and with all our mind and with all our strength. AI can never do this, for the simple reason that to love the Lord with mind, hand, heart, and soul is, uniquely, part of what it means to be human.