A machine can now compose a prayer. It can imitate the cadence of the Psalms, borrow the language of repentance, and weave together words of humility, mercy, and grace. It can invoke the name of God, quote the Holy Scriptures, summarize the teachings of the Holy Fathers, and produce a sermon in less time than it takes for a monk to complete a single Jesus Prayer. Is there something the machine cannot imitate?
What it cannot do is repent. What it cannot do is weep. What it cannot do is stand, like the Publican, beating its breast and crying from the depths of existence: “God, be merciful to me, a sinner.”
That distinction may sound obvious. In the age of artificial intelligence, it is becoming less so.
The public debate over AI is usually framed in practical terms. Will it take jobs? Will it transform education? Will it flood the internet with synthetic writing? Will it make us more productive or more dependent? These are urgent questions. But beneath them lies a deeper one: What exactly are we comparing the machine to?
If the human mind is only a system for processing information, then artificial intelligence looks like a rival. If language is nothing more than a statistical arrangement, then a machine that produces fluent language begins to look disturbingly human. If intelligence is measured primarily by output, the difference between a person and a program begins to blur. Yet this blurring reveals less about the machine than about the poverty of our anthropology.
The Orthodox Church does not understand the human being as an intelligent mechanism, but as a living icon of the living God. This is where Gregory Palamas, a 14th-century Byzantine monk and one of the central theologians of Orthodox Christianity, becomes unexpectedly relevant. Palamas defended the hesychast tradition, a spiritual discipline of prayer, silence, and ascetic struggle. At the heart of that tradition was the conviction that human beings can truly participate in God’s grace through the purification of the heart, the illumination of the nous, and communion with the divine energies, even though God’s essence remains beyond human comprehension.
This distinction between the divine essence and the divine energies is central in Palamite theology. “Divine energies” does not mean energy in the physical sense. God in His essence remains utterly unapproachable, beyond every concept and every created faculty. Yet in His energies, He truly gives Himself, becoming present and communicable without ceasing to be incomprehensible. Human beings do not possess God as an object of analysis. They encounter Him, by grace, through transformation.
This is a very different account of knowledge from the one implicit in much of our technological culture.
Palamas uses the Greek term nous, often inadequately translated as “mind” or “intellect.” In Orthodox theology, the nous is not simply the part of us that calculates, argues, or forms sentences. It is the deepest faculty of spiritual perception, the place where the human person becomes capable of knowing God.
Reason is not rejected. Orthodoxy has always honored the sanctified intellect. But reason is not the whole of man, refusing to treat thought as the whole of the person.
Artificial intelligence, by contrast, works through formalization. Language, images, sounds, and decisions must be converted into data, tokens, categories, and patterns. A system can then detect relationships, predict what comes next, and generate results that often appear intelligent. Something real is captured in that process. Much is also left outside it.
This is not a defect in AI. It is the condition of its usefulness. A tool must reduce reality in order to work with it (sometimes this process is called linearization). A map leaves out nearly everything except what the traveler needs. A medical chart reduces a patient to numbers, symptoms, and history. A translation reduces one language into another imperfectly, because no language carries another whole.
Something real is captured in this process. Yet something essential remains forever beyond its reach.
The problem begins when the reduction is mistaken for the whole.
AI can be immensely useful. It can help researchers sort evidence, doctors detect patterns, teachers prepare material, journalists search archives, and students organize their thoughts. Used properly, it is not an enemy of human intelligence. It is an instrument, a very smart and helpful tool.
But an instrument is not a self. And a system that can generate religious language has not thereby entered the religious life.
A chatbot can write, “Lord, have mercy.” It cannot need mercy. It can describe fasting. It cannot hunger for purification. It can explain repentance. It cannot stand ashamed before another person, or before God, and ask to be changed.
This is why the current debate over AI touches, whether we admit it or not, on the older debates over transhumanism and posthumanism. Transhumanism imagines technology as a path to overcoming biological limits. Posthumanism questions whether the human being should remain the central measure of meaning at all. Both can ask useful questions. But both also expose a temptation that is now everywhere: to treat the person as an upgradeable system.
Once this assumption is embraced, artificial intelligence appears less as an instrument and more as the next evolutionary stage.
Palamas would have seen the danger in different terms. For him, the person is not defined by output, efficiency or problem solving. The human being is created in the image and likeness of God and called to theosis. The nous is not merely an advanced cognitive function. It is the faculty through which the person is opened to divine life.
One of the more striking parts of Palamas’ thought is his distinction between levels of inner discursive reason (logos). There is the ordinary inner discursive reason, the movement of thought that can become argument, calculation or spoken language. But there is also a deeper Word rooted in the nous, a form of knowledge inseparable from the person’s spiritual life.
In simpler terms, Palamas distinguishes between the kind of thought that can be expressed and arranged, and the kind of spiritual knowing that belongs to a person’s relationship with God.
AI operates in the first realm. It has no access to the second. That is not because we are not yet technologically advanced enough.
This truth should not lead us to fear technology or to reject its legitimate uses. Machines do not need souls to be helpful. A calculator does not understand mathematics, but solves difficult problems. Large language models do not love poetry, but can produce poems based on the poet’s statistical writing. Likewise, AI need not be conscious to reshape the world.
But we should be precise about what kind of change this is.
The history of mathematics offers a valuable lesson. Kurt Gödel demonstrated through his incompleteness theorems that no formal system can contain all truth within itself. He reminds us that reality always exceeds the structures we build to describe it, showing that formal systems have limits that cannot be overcome simply by making them more elaborate.
Gödel should not be dragged into the AI debate as a magic proof that machines will never think. That would be careless. But he remains a warning against intellectual arrogance, especially the fantasy that all truth can be contained within a closed formal structure, as in AI models.
The Palamite warning is somehow different, but it points in a related direction. The deepest truths of the human person are not captured simply by description and statistical formalization. Love can be described. So can grief, courage, shame, prayer, and death. But description is not possession.
A machine can produce a moving paragraph about a mother burying a child. It has no experience of losing a child. It can write about forgiveness. It has no experience of the moment of forgiveness. It can produce a theological essay on divine light. It has not stood in silence before mystery. Such knowledge is not vague. It is too real to be exhausted by description and explanation.
For Palamas, the highest knowledge is not irrational, not against reason, but beyond what reason alone can possess. The Christian mystical tradition has always known that some truths can be approached only through transformation. Apostle Paul gestures toward this when he speaks of hearing things that could not be fully uttered in ordinary human speech.
Our age is tempted to believe that whatever cannot be formalized is secondary, or even nonexistent. We trust what is measurable, scalable, and repeatable. We increasingly accept what can be processed as more real than what can just be lived. Artificial intelligence did not create this temptation. It inherited it and made it dazzling.
This is the real spiritual risk. Not that machines will suddenly acquire souls, but that human beings will forget they have them.
Religious communities should not respond to AI with panic. They should also not baptize every new technology as progress. The proper response is discernment (διάκρισις). AI can serve human beings when it remains ordered toward human purposes. It becomes dangerous when it begins to shape our understanding of truth, wisdom, and personhood in accordance with its own formal limitations.
The question, then, is not whether artificial intelligence will become more impressive. For sure, it will. It will write better, speak better, diagnose better, and imitate better. In many fields, it will outperform us.
The crucial question is whether we will still be able to recognize that personhood is a mystery of being rather than a matter of performance.
A machine may learn our language. It may even learn our theology, tuning up even the Orthodox Theology, with patristic texts as the training set. But it cannot enter the silence from which prayer begins.
Protopresbyter of the Ecumenical Throne Gregory-Telemachos Stamkopoulos is a clergyman of the Greek Orthodox Archdiocese of America, theologian, and scholar of applied information. He is an Associate Professor at Aristotle University of Thessaloniki and the director of the Lab of Digital Innovation in Theology and Culture, where he teaches Science, Information, and Orthodox Theology. His research focuses on artificial intelligence, neural networks, Orthodox theology, and the philosophical dimensions of information.
