本文探讨了天主教会关于人工智能的立场，重点关注其与人类本质的四个关键特征——理性、真理、肉身性和关系性之间的关系。文章分析了教会的观点，认为人工智能在这些方面与人类存在本质上的差异。作者深入探讨了教会对这些概念的理解，特别是在肉身性和关系性方面，并将其置于更广泛的哲学背景下进行讨论。文章还提及了教会对超人类主义的看法，以及其对传统道德框架的影响。

🤔 天主教会认为人工智能在理性、追求真理、肉身性和关系性这四个方面与人类存在本质差异，构成了其核心观点。

🧠 在肉身性与关系性方面，文章探讨了“肉身”的重要性，强调身体在定义身份和道德方面的重要性，这与“理性主义”的观点有所不同。

🤝 关于关系性，文章阐述了道德义务源于人与人之间的关系，而非个体本身。这与人工智能和外星生命体的道德考量相关，并探讨了天主教会在这一框架下的立场。

⛪ 天主教会在人工智能问题上，试图在古典希腊二元论和第一世纪犹太一元论之间找到平衡，这反映在教会对“灵魂”和“身体”关系的看法上。

Published on May 19, 2025 1:30 PM GMT

On January 28, 2025 (during the pontificate of Pope Francis) the Catholic Church put out its position paper on AI: Antiqua et nova: Note on the Relationship Between Artificial Intelligence and Human Intelligence. Here are my thoughts, which I wrote back in February, but not much has changed since then, and the new pope has continued citing AI as of particular concern to the Catholic Church.

---

"With all your modern science, are you any closer to understanding the mystery of how a robot walks or talks?" "Yes, you idiot. The circuit diagram is right here on the inside of your case!" (Futurama)

Four characteristics of humanity

Section III is the real meat of the document. It lists four essential characteristics, and argues that AI falls short of the fullness of humanity in each respect. These are: rationality, truth-seeking, embodiment, and relationality.

I will address embodiment and relationality first, since these two things are connected, and they are in my opinion the most interesting ones. By comparison, I understand less of what (if anything) the document has to say about rationality and truth-seeking, so I'll address those briefly afterward.

Embodiment and relationality

Lately I've been on a contrarian kick advocating these stances in opposition to the "modal rationalist" views of disembodiment and individuality. However, if I take off the contrarian hat, I suppose my actual position is something like:

The general view of rationalists (including myself in the past) is too far biased towards the extreme of disembodiment and individuality because of certain philosophical premises which are usually left unexamined, and which are not empirically-based. The correct position is probably somewhere more in the embodiment/relationality direction on the spectrum.As a purely descriptive matter, embodiment and relationality seem to more accurately explain the way people actually behave, as opposed to modal rationalism. Understanding this is important for living in society. It also poses a conundrum for attempts to define a "utility function (CEV) of humanity" - if most humans are not utilitarians, how can there be such a thing?

What are embodiment/disembodiment and relationality/individuality? It is helpful to define these positions ostensively by giving examples of the views on the extremes of these two spectra.

Disembodiment:
What matters (for morality and personal identity) is a certain pattern of information that constitutes the experience of a conscious being.If that pattern were to be implemented on some other substrate (e.g. mind uploading), then that implementation is (again, for the purpose of morality and identity) equivalent to a flesh-and-blood person.In humans, that pattern resides in the brain; the rest of the body doesn't matter.
This is an empirical claim that doesn't necessarily follow from the other premises, but I think those other premises give rise to a bias that leads people to just assume that this is true without any empirical basis. I think the correct view is: Yes, the brain probably has most of the stuff-that-matters, but the rest of the body still has a substantial portion, more than most rationalists give it credit for. Identity is not all-or-nothing, but operates in a "holographic" manner. Just as people who experience brain damage end up changing to some extent, depending on the extent of the damage, so do people change when their bodies change. A complete replacement of the body-below-the-neck (as in e.g. brain-only cryopreservation) would probably change a lot as well.
Individuality:
The moral value of a thing rests on characteristics of the thing itself.If we were to discover intelligent aliens, our moral obligations towards them would be the same as to other humans.If we develop a sentient AI, then we also ought to treat it the same way we treat humans.Given that babies, mentally impaired humans, and certain species of animals all have equivalent levels of intelligence, we must treat them all the same in order to be consistent.
Embodiment:
What matters is the physical substance of which a being consists.
I am not sure I would endorse this fundamentally, but I would at least say it is "more true than not-true for most practical purposes". That is, I still buy the argument that identity isn't in specific atoms, and I accept the metaphysical possibility of my identity being copied multiple times, but that doesn't imply that copying a mind is an easily-achievable feat of engineering like copying a computer file, or that swapping out parts of my body or brain with entirely different materials (e.g. silicon transistors) would not have a significant effect on my identity.
A reimplementation or upload of a human onto a non-biological substrate would not matter at all.
Relationality:
One being has moral obligations to another depending on the relationship between them. There is no such thing as "moral value" per se.These relationships are not necessarily based on voluntary agreement (contracts), since there are some that exist prior to any deliberate act. For example, a newborn child has a morally relevant relationship with his/her parents, even though he/she could not possibly have agreed to it. Similarly, the citizens of a particular nation have a stronger relationship to each other than to foreigners, even if they didn't choose to be born there.If we discover intelligent aliens, we would have no moral obligations towards them (and likewise they would have none towards us).

(Why are embodiment and relationality connected? Because a combination of both can be used to derive a "universal" (or rather "all-human-" or "all-Earth-including") morality based on the physical connection all humanity (or, as some environmentalists would say, all life on earth) shares via common descent. But then this stance diverges radically from the modal rationalist view when it comes to AI/uploads and aliens.)

Relational morality can be rather harsh by modern standards, as we see from the way it was lived out in classical antiquity:

Two warring tribes or nations have no obligations towards each other. Each only has an obligation to try to gain as much territory and resources for their own people as they can.A nation can only be unified by a (perhaps legendary) common line of descent.The paterfamilias has absolute power over the rest of his family; if he decides that an infant should be exposed, then that's what happens.Slaves are not part of society and so they have no rights in that society.People believe that the gods treat humans as their playthings, and they are considered not as moral exemplars, but as capricious forces that must be placated through sacrifices. (This ethos of classical paganism is summarized in the saying do ut des - "I give so that you may give".)

Now, we can understand the Catholic Church's position (and perhaps that of classical monotheism generally) as a way of mitigating this harshness by introducing a single "God", while still remaining within the relational framework because universal morality is founded upon everyone's relationship with God. Since all humans have that relationship, they still have rights and obligations vis-à-vis each other by way of God, even if no social relationship between them exists.

(Incidentally, I think this helps explain what's going on in the "moral argument for the existence of God", which, if considered apart from this historical background, it seems puzzling why anyone would consider compelling. Why do some people insist that God must exist in order to "ground" morality? What does "grounding" even mean here anyway? But now we see that this insistence makes sense if we're thinking of morality relationally. In that framework, since moral consideration is a characteristic of relationships and not of individuals, then for any universal morality to exist, then there must be some entity that universally has a relationship with everyone, i.e. God.)

Similarly, the full "embodiment" perspective is also hard to face because of its obvious implication that there can be no afterlife, and the pain and suffering in the world seems to have no purpose. Some religions/philosophies have reacted to this by going to the opposite extreme (Manicheanism/Gnosticism) where the material world is a bad thing and our goal should be to escape it. The Catholic Church also rejects this view, and likens transhumanism to it (endnote 9).

However, it remains unclear to me what their actual stance is. If "the entire human person is simultaneously both material and spiritual" (paragraph 16), then what is it that goes to heaven or hell after death? As far as I can tell, the view of the historical Jesus (and the general current of Jewish thought at the time) was that the "soul" does not exist apart from the body, and so at the end of time, God or the Messiah would physically re-create the bodies of all people who have died so they can live eternally in the Kingdom of God (which would be a physical kingdom here on earth). The obvious consequence of this monistic metaphysics is the doctrine of "soul sleep" whereby people stop existing at death and only come back when God resurrects them at the end of time. However, the Catholic Church rejects that doctrine as well. At any rate, it seems that, whereas their position on the relational/individual axis is still fundamentally relational (just with God added), their position on embodiment/disembodiment is more middling - "kinda disembodied, but not fully" - reflecting a synthesis between classical Greek dualism and first-century-Jewish monism.

What does this all have to do with AI? Antiqua et nova says that AI is totally and essentially different from humans because it lacks all the morally-significant characteristics that humans have:

AI does not have a body.To the extent that there's some non-material aspect of humans ("spirit"), AI lacks that as well, because it was not made by God in his image.AI has no relationship with God.AI also has no relationships with humans in the way that matters (see paragraph 61 and endnote 123).

Setting aside the points about God, let's consider points 1 and 4. This is what I think about those:

This is definitely true about current AI.This is probably going to continue to be true given the current trajectory of AI development.
If AIs are built with robot bodies, they would not be anything like human bodies. For example, they would not be made of biological matter, and they will not be integrated with the intelligence in the way the human body is, but rather separated by a "quasi-Cartesian" divide.AIs are not going to develop attachments or friendships with humans.
This is not necessarily always going to be the case (that AI will lack humanlike embodiment and relationality), but achieving this would require a concerted effort, and nobody seems to really be working on this.If the arguments made in these sections of Antiqua et nova are taken more seriously, then perhaps people interested in making human-aligned AI will prioritize these research efforts more, since embodiment and relationality seem to be important parts of human morality.Perhaps a more viable route towards that goal would be to use AI to augment the capabilities of existing humans (while maintaining their embodiment and relationality), rather than building something from scratch.

Rationality

Rationality is cited as another essential characteristic of humanity, what makes us distinct from animals. On its face, this would seem to be one of the easiest aspects for AI to replicate; however, the document claims that the capabilities of machines are only a small subset of what humans do with their intelligence (paragraph 27). But the examples given there seem tenuous; there is no real justification given for why we should not expect AI to soon achieve those abilities as well - unless we artificially circumscribe the definition of "intelligence" etc. in such a way as to exclude AI, e.g. "AI's advanced features give it sophisticated abilities to perform tasks, but not the ability to think" (paragraph 12). This resembles John Searle's view (Chinese room, etc.) that what a machine is doing is not "really" understanding, cognition, etc. I think this view is going to be quickly rendered obsolete by advances in AI.

Anticipating this, the document also makes sure to reject altogether the position it calls "functionalism" (which I think is a confusing word to use; "behaviorism" is the word more consistent with the philosophy-of-mind literature), that what an AI or a human can do is what matters (see paragraph 34). This betrays a lack of confidence in the point made immediately prior (paragraphs 32 and 33) which claims (unjustifiably, in my opinion) that there are certain capabilities that are fundamentally out-of-reach for AI: "Since AI lacks the richness of corporeality, relationality, and the openness of the human heart to truth and goodness, its capacities—though seemingly limitless—are incomparable with the human ability to grasp reality" (paragraph 33) - but I think it's only a matter of time before AI has all of that "richness".

Much of the document's confidence in the distinctness of humanity rests on the notion that mankind is created in the image of God (Genesis 1:27). But the document also quotes Augustine to clarify what this means: "Man is made in the image of God in relation to that [faculty] by which he is superior to the irrational animals. Now, this [faculty] is reason itself, or the 'mind,' or 'intelligence,' whatever other name it may more suitably be given" (endnote 16, brackets in original). This seems to leave the Church an escape-hatch in the event that AI capabilities improve to equal those of humans in the relevant aspects - they can say that the AIs now possess the divine image as well!

Truth-seeking

Furthermore, according to the document, human intellect is "irresistibly" drawn towards truth (paragraph 21). Now, in the context of AI, one might say that there are certain convergent instrumental goals that every intelligent agent will take on, among which is truth-seeking, i.e. the desire to have a more accurate world-model, since this is useful as a means towards almost any goal that the agent might have in interacting with the world. However, that doesn't seem to be what the document means by "truth". It refers to "realities that surpass mere sensory experience or utility" (paragraph 20) and "realities that transcend the physical and created world" (paragraph 23), ultimately equating Truth with God. This is distinct from the instrumental goal of world-modelling (but of course, the very concept contains the embedded assumption that there are indeed any such "realities", which many would dispute). The document also suggests (paragraph 22; endnote 40) that "semantic understanding and creativity" are examples of this transcendent faculty, with the implication that an AI cannot replicate them. However, the possibility for computers to do these things has been well explored even prior to the LLM paradigm (e.g. in Gödel Escher Bach), and, again, I think current AI is not far from achieving it even if it remains limited to the kind of AIs (lacking embodiment and relationality) that currently exist.

How much of Antiqua et nova is specifically theistic, or Christian?

If you don't think that God exists and/or don't think Jesus was a divine being who rose from the dead (etc.) you may be wondering how much of this document is relevant, or whether one must first accept the truth of those things before finding anything useful in it.

The document is certainly not "rationalist" or "scientific" in its approach, but at the same time I was surprised by how little doctrinaire Christianity is present in it. (But maybe this is true of Catholic thinking generally, and my surprise is due to my being more familiar with Protestantism.) God is mentioned throughout, but as a highly abstract and rarefied concept (the God of the philosophers, as opposed to the God of Abraham) to which the likes of Plato, Aristotle, or even the Enlightenment-era deists would have little objection. Jesus is only mentioned a few times.

What kind of theism is the document based on?

Continuing the above discussion of "truth-seeking", we see that the document largely conceptualizes God as some kind of supreme, transcendent principle towards which all humanity is (or ought to be - there is a tendency to equivocate between is/ought) striving (paragraph 23). Now, I find this concept rather vacuous and needlessly mystifying, a typical "iatrogenic" problem created by philosophers in order to create more work for other philosophers.

There are ways to set up a similar concept non-theistically, by regarding truth, philosophy, mathematics, etc. as something worth striving for for non-utilitarian reasons. E.g. "It appeared to me that the dignity of which human existence is capable is not attainable by devotion to the mechanism of life, and that unless the contemplation of eternal things is preserved, mankind will become no better than well-fed pigs" (Bertrand Russell, Autobiography). However, this seems like a very niche interest of a peculiar sort of person. More generally, whether theistic or not, the idea of searching for an eternal/transcendent/supreme/etc. principle seems far more prevalent in the Western philosophical tradition than elsewhere, and so it's a stretch to regard it as the very essence of human intellect. (Admittedly, I am not as familiar with non-Western philosophy, so I may be wrong.)

Besides that, what specifically Christian points are made?

The "Stewardship of the World" section cites the Book of Genesis to say that humanity has a special position of authority over the earth, granted by God, and so although were are not called to merely live in harmony with nature (as "secular Gaianists" would say), we are also not supposed to remake the world in the likeness of humanity (as would be the view of secular humanists, techno-utopians, accelerationists, etc.). Rather, according to the Christian view, we have an obligation to use our technological mastery to better glorify and achieve union with God.

It's not clear what this actually means in concrete terms. The document takes pains to emphasize that the Church doesn't oppose technological progress per se (e.g. paragraph 2), but this "stewardship" doctrine leaves them a lot of leeway to define what exactly constitutes "godly" or "ungodly" progress. This will certainly be a source of debate in the years ahead.

The Christian doctrine of the incarnation (that God became flesh in the person of Jesus Christ; see paragraph 16) is mentioned in support of the importance of embodiment. However, embodiment seems largely orthogonal to Christianity - there are/were flesh-disparaging forms of Christianity (Gnosticism, and perhaps some forms of Protestantism) and of atheism (transhumanism). At the same time, I would argue (though perhaps some Christians and rationalists will alike dispute this) that embodiment is the natural consequence of the kind of monistic materialism that most rationalists claim to espouse, even though they may end up trying to reconstruct an ersatz form of Cartesian dualism in order to shore up a structure of philosophical concepts that was originally built on a dualistic foundation.

Lastly, unless I missed something, the only specific mention of the teachings of Jesus is in paragraph 20: "Love of God cannot be separated from love for one’s neighbor. By the grace of sharing God’s life, Christians are also called to imitate Christ’s outpouring gift by following his command to 'love one another, as I have loved you'" (citations omitted).

Thus, although (as I have argued) the idea of relationality is not peculiar to Christianity, the document justifies its particular stance that the moral dimension of humans' relationship with God (and, by implication, with each other) ought to be one of service and self-sacrificing love (as opposed to e.g. obedience, loyalty, propriety, duty, etc.) by referring to Jesus's commandment and the Christian belief that he sacrificed himself for the rest of humanity.

Discuss