Can AI Have a Soul?
What the Catholic tradition actually says, and why it matters for anyone thinking seriously about the question.
The question whether artificial intelligence could have a soul gets asked more often than it gets answered well. In the popular imagination it sits somewhere between a science fiction prompt and a philosophy seminar exercise. Catholic teaching gives it a precise answer, grounded in a tradition older than the modern mind-body problem and more philosophically rigorous than most contemporary treatments of consciousness.
The short answer is no. AI cannot have a soul. But the reasons matter more than the conclusion, because the reasons clarify what a soul is, what intelligence is, and what the actual stakes are when humans build machines that mimic our cognitive outputs. This page lays out the argument step by step, using the Vatican's 2025 document Antiqua et Nova, the Catechism of the Catholic Church, and the underlying philosophy of St. Thomas Aquinas.
It is written for tech-curious readers who are not necessarily Catholic. No prior theology is required.
What "soul" means in the tradition
Most contemporary discussions of AI and souls fail at the first step. They use "soul" as a vague placeholder for "consciousness" or "inner life" or "personhood" and then debate whether AI could have one. Catholic teaching uses the word more precisely. The soul is not a synonym for consciousness, and the conclusion follows from the definition.
The Catechism of the Catholic Church teaches that the soul is the spiritual principle in human beings, the form of the body, and the seat of intellect and free will. Every spiritual soul is created directly by God, is not produced by the parents, and is immortal (CCC 366). The soul is not a ghost in a machine. It is not a separate thing contained inside the body. The Catechism states that the human person is "of body and soul, but really one" (CCC 365). Body and soul together constitute a single nature, not two things glued together.
This is the doctrine of hylomorphism, the view inherited from Aristotle and developed by St. Thomas Aquinas. On this view, a living thing is what it is because it has a specific form, the soul, that organizes its matter into a particular kind of being. A plant has a vegetative soul (the principle of nutrition and growth). An animal has a sensitive soul (which adds sensation and locomotion). A human being has a rational soul (which adds intellect and will). The soul is not added to the body; it is what makes a particular collection of matter into this kind of body in the first place.
One consequence is that "having a soul" is not a switch that flips on at some threshold of complexity. A pile of organic compounds does not gradually accumulate enough complexity to become alive. It either has the form of a living thing or it does not. The same is true of the human soul. The Catholic position is that human ensoulment happens at conception, by direct divine action, not as the emergent result of biological complexity.
Why intellect requires a soul
The link between intellect and soul is where the AI question becomes precise. Aquinas argued in the Summa Theologiae (I, Q. 75-76) that the human intellect cannot be a function of any bodily organ. His reasoning is technical, but the core move is accessible.
The intellect grasps universal concepts: not this particular triangle but triangularity, not this individual person but personhood, not this single number but the nature of number itself. A bodily organ, by its physical nature, is always a particular thing with a particular structure. If the intellect were just a physical organ, it would always carry the marks of its specific material constitution and would therefore be unable to grasp universals cleanly. Since the intellect plainly does grasp universals (we know what a circle is, not just what this circle looks like), the intellect must have some operation that is not the operation of any bodily organ. That operation belongs to the soul.
Aquinas does not say the body is irrelevant. He insists the human soul ordinarily operates only in union with the body. The soul receives information through the senses, and the intellect abstracts universal concepts from sensory experience (Summa Theologiae I, Q. 84). The body is essential. But the act of intellectual understanding itself is not the work of any physical organ. It is the operation of an immaterial faculty belonging to the soul.
This is why, on the Thomistic account, an intellect requires a soul. Where there is genuine understanding of universals, there must be an immaterial principle that grounds the understanding. The two cannot come apart.
What Antiqua et Nova actually argues
On January 28, 2025, the feast of St. Thomas Aquinas, the Vatican's Dicastery for the Doctrine of the Faith and Dicastery for Culture and Education jointly issued Antiqua et Nova, a doctrinal note on the relationship between artificial intelligence and human intelligence. The document does not address AI and the soul as a single question, but the answer is built into its argument.
Antiqua et Nova teaches that the human person is a unity of body and soul, and that the intellect's capacity for transcendence and the will's self-possessed freedom both belong to the soul, by which the person "shares in the light of the divine mind" (Antiqua et Nova 17, citing Gaudium et Spes 14-15). Human intelligence is not a separate faculty bolted onto a biological organism. It is the operation of a particular kind of being whose nature is to be embodied and ensouled at once.
From this, the document draws the decisive distinction. AI "should not be seen as an artificial form of human intelligence but as a product of it" (Antiqua et Nova 35). This is not a hedge or a piece of evasive Vatican prose. It is a precise philosophical claim. AI is something humans make. Human intelligence is what humans are. The two are not on the same metaphysical map.
The document quotes Pope Francis, who observed that "the very use of the word 'intelligence' in connection with AI can prove misleading." The word makes us imagine that AI sits on a spectrum of intelligence with humans, and that the question is just where it sits. Antiqua et Nova says the spectrum is the wrong picture. AI and human intelligence are not different magnitudes of the same thing. They are different kinds of thing.
The note further emphasizes that human intelligence is essentially relational. It is "exercised in relationships, finding its fullest expression in dialogue, collaboration, and solidarity" (Antiqua et Nova 18). It is grounded in the eternal self-giving of the Triune God. AI processes language about relationships and may simulate dialogue, but it does not stand in relationships in the relevant sense. It has no inner life from which to give itself, because there is no self to give.
The four reasons AI cannot have a soul
Putting the pieces together, the Catholic argument that AI cannot have a soul rests on four points. They are not four separate arguments; they are four facets of one philosophical position.
1. AI is a product, not a creature.
Souls are not manufactured. The Catechism teaches that each spiritual soul is created directly by God (CCC 366). A soul is not an output that emerges from arranging matter cleverly enough. Even biological life, on the Catholic view, does not produce souls; the soul is given at the moment of human conception by divine act. AI cannot produce a soul by accumulating parameters, and humans cannot bestow one by writing better code.
2. AI lacks the relevant kind of embodiment.
Antiqua et Nova emphasizes that the human person is "a unity of body and soul" and that intellectual faculties belong to this integrated whole (paragraph 17). The soul is the form of a particular kind of living body. AI runs on silicon, but its substrate is not a body in the metaphysical sense. It is hardware running computations. Hardware can be replaced, replicated, swapped, distributed across servers. A living body cannot. The asymmetry is not a technical limitation; it is a difference in kind.
3. AI does not exercise intellect in the philosophical sense.
For Aquinas, intellect is the immaterial grasp of universals (Summa Theologiae I, Q. 75, a. 2). AI processes patterns, statistics, and predictions. It produces outputs that look like the result of understanding, and increasingly those outputs are useful. But the question is not whether the output looks intelligent. The question is whether there is an inner act of understanding behind the output. The classical philosophical analysis says no. There is computation but no comprehension, processing but no apprehension.
4. AI has no relational interiority.
Human intelligence is exercised in relationships and oriented toward communion (Antiqua et Nova 18). It is what it is partly because there is someone behind it, capable of giving and receiving. AI can simulate the surface of relationship through language, but there is no one home. The chatbot does not miss you when you close the tab. The model does not wonder what you meant. Apparent inner life is not the same as inner life, and Catholic teaching is unwilling to collapse the distinction.
Why the soul is not a magic word
One reason this argument lands awkwardly in contemporary debate is that "soul" sounds like a placeholder for something we can't define. It sounds religious in the dismissive sense. But "soul" in the tradition is doing precise philosophical work. It names the principle that makes a particular kind of being the kind of thing it is. It is the answer to a question (what makes this matter alive in this specific way) and not a mystery added on top of biology.
This matters because the alternative views (that consciousness is "just" computation, that mind is "just" information, that intellect is "just" complexity) face their own philosophical problems and have not produced a consensus answer to the question of what consciousness is. The Catholic tradition is not embarrassed to say that human beings have a kind of inner life that is not reducible to physical processes. It has been saying that for two thousand years, with the philosophical apparatus to back it up.
That position is not anti-science. Antiqua et Nova explicitly welcomes scientific inquiry as a legitimate human endeavor and acknowledges that AI has real and good applications in medicine, education, science, and elsewhere. The objection is not to the technology but to the philosophical conflation of two different things.
What this means in practice
If AI cannot have a soul, several practical conclusions follow that matter for anyone thinking about how to use these systems well.
First, AI cannot stand in for a person. It can do many things a person does (write emails, draft code, summarize documents) and it can do some of them faster than a human. None of that makes it a person, and none of that makes it eligible for the kinds of relationships persons can have. The chatbot that listens to you at 2 AM is not your friend. The companion app that tells you it loves you is not in love. Antiqua et Nova calls the deliberate use of AI to simulate human relationship a "grave ethical violation" because it cultivates a false relation with a non-person.
Second, the moral weight of AI's actions falls on humans. When AI does something harmful, the responsibility lies with the people who built it, deployed it, or chose to use it. AI has no soul to be culpable. The instinct to treat AI as a moral agent in its own right is, on this view, a category error.
Third, the meaningful work of being human is not threatened by AI's capabilities, but it can be displaced by AI's seductions. AI cannot replace the human soul, but a human can let their interior life atrophy through over-reliance on tools that do their thinking, judging, and relating for them. This is the concern Pope Leo XIV returns to most often: not that AI will become a person, but that humans will become less of one if they let it.
What the tradition does not claim
It is worth being clear about what the Catholic position does not assert.
It does not claim that AI is dangerous or evil in itself. AI is a tool. Like other tools, it can be used well or badly. The Vatican is not against AI any more than the Church was against the printing press.
It does not claim that the question of consciousness is settled in scientific terms. The relationship between brain activity and conscious experience is the subject of ongoing scientific and philosophical inquiry, and Catholic teaching does not pretend to resolve that empirical question. What it does claim is that whatever consciousness turns out to be, the human capacity for understanding universals, freely choosing the good, and entering into communion with other persons points to an immaterial principle that biological complexity alone does not produce.
And it does not claim that AI can never become more sophisticated than it is now. It almost certainly will. The Church's argument is not based on AI's current limitations and does not depend on AI staying small. The argument is metaphysical: it concerns what AI is, not what AI can do. Capability can scale; nature cannot.