What the Vatican gets right, and wrong, about AI

The Oxyrhynchus Papyri, discovered in rubbish mounds along the River Nile during the era of British governance in Egypt, offer a fascinating cross-section of documents and correspondence across nearly a millennium of history, from the Ptolemaic era to the end of Roman rule. In one of them, a man named Hilarion writes to his pregnant wife Alis. ‘[I]f you bear a child and it is male, let it be… [but] if it is female, cast it out.’ With that, Hilarion, away in distant Alexandria, instructs his wife to kill whatever daughter she may have.

What would now be documentary evidence for murder in the first degree was a plain commonplace back then. The reasoning for the murder of Alis’ baby girl is unstated in Hilarion’s letter, but it is known from practice throughout the world of Roman antiquity. Daughters were cost centres, requiring upkeep and dowry, and management as property. Sons were profit centres, with all that implied. A low-status couple like Hilarion and Alis could ill afford a daughter, and so any who arrived were simply to be ‘cast out’, usually subjected to an ordeal known as exposure, in which the helpless infant was abandoned in the open, dying in time of sun, weather, thirst, hunger and neglect. The steady supply of infants thus abandoned was coincidentally a ready source of slaves for merchants collecting and selling their fellow human beings. 

This practice, in its callousness and cruelty, depended fundamentally upon a view of the human person as a commodity, devoid of intrinsic value, and justified in his life only to the extent that he – or she – provided that justification through material profit, production or delight in others. There was no right to live, and there was no concept that a human life might be an end in itself, surpassing all other ends in worth. 

It is this debased idea of man, an idea that has come roaring back in modernity, as a disposable thing rather than the very Image of God, that moved the recent issuance of Magnifica Humanitas, the first encyclical of the reign of Pope Leo XIV. The Papal contention, and that of the Papacy before this Pope, is that the advance of artificial intelligence is a prime catalyst (though not a sole catalyst) for the revival of that prior framework, and that it must be understood and fought in the context of the past centuries of Christian understanding of what man is – and what man is for. 

Preceding Magnifica Humanitas is a significant corpus of Papal and Magisterial teaching on the anthropology of man versus the challenges of modernity. Those works include, but are not limited to, Rerum Novarum, issued 15 May 1891 under Pope Leo XIII; Quadragesimo Anno, issued 15 May 1931 under Pope Pius XI; Nostra Aetate, issued 28 October 1965 under Pope Paul VI; Humanae Vitae, issued 25 July 1968 also under Pope Paul VI; Spe Salvi, issued 30 November 2007 under Pope Benedict XVI; Dignitas Infinita, issued 8 April 2024 under Pope Francis; Antiqua et Nova, issued 28 January 2025 also under Pope Francis; and Quo Vadis Humanitas, issued 4 March 2026 under Pope Leo XIV as a precursor to his first encyclical. Not all of these prior works are encyclicals. Dignitas Infinita is a declaration of the Roman Curia’s Dicastery for the Doctrine of the Faith; Antiqua et Nova is a joint note of that Dicastery and the Dicastery for Culture and Education; and Quo Vadis Humanitas is an International Theological Commission text. 

It should also be noted that this isn’t even a comprehensive list of relevant source material: Pope St John Paul II’s 1991 Centesimus Annus, for example, might be included, and there is more in that vein. Nevertheless, in labouring towards a grasp of the Catholic Church’s understanding of man and his place in the cosmos, these documents give a fairly good overview of it. 

Taken altogether, they are rich and consistent in their teaching. (One might single out Humanae Vitae and Quo Vadis Humanitas as especially indispensable.) Across time and sources the Magisterium teaches that man alone bears the Imago Dei; that man is both body and mind, inseparable; that the fact of the Incarnation honours man above all other intelligent creations; that the nature of man’s intellect is irreproducible; and that all efforts to supersede, overcome or replace man are both doomed and intrinsically evil. All this is expounded and powerfully defended in these works, and also in the deep vein of works and thought upon which they themselves are grounded. St Thomas Aquinas gives us the understanding of rational man as a directed, not a reactive, intelligence, that seeks the good; Aristotle gives us the understanding of man’s communities as relationally oriented toward a telos, a directed end, that is ideally the good; and Christianity itself gives us the Trinitarian God, whose interiority of loving relations in Persons establishes the template for man himself in his relations to others. 

The rootedness of this grasp of man predates modernity, even if modernity calls it to fresh urgency. Dante, in Inferno’s Canto XXVI, has Odysseus in hell remind the voyagers of their intrinsic nature:

Considerate la vostra semenza:

fatti non foste a viver come bruti,

ma per seguir virtute e canoscenza.

Consider ye the seed from which ye sprang;

Ye were not made to live like unto brutes,

But for pursuit of virtue and of knowledge.

Artificial intelligence is capable of none of this and never will be. As document after document across the past decade points out, it is not embodied, it is not wise, it is not experiential, it is not moved by love or grief or compassion or sorrow or mercy, and it is above all else not created in the Imago Dei. 

If the good people of Oxyrhynchus murdered their daughters on the mistaken belief that utility was the measure of man, then they would not be mistaken to think it the measure of the mimicry of man’s mind in AI. Utility, the Vatican reminds us again and again, is the telos of AI, and that utility must be assessed according to the telos of man. Does it direct us to the good? Does it facilitate the virtues? Does it enhance or diminish our human dignity – in the categories of Dignitas Infinita – in ontology, in society, in morality, and in existence? Underlying the Catholic critique of AI is that the answer is too often ‘no’. 

The expectations around Magnifica Humanitas were that it would set a capstone upon this prior work, unifying and directing Catholic thought and teaching on non-human intelligence in the same manner as did Rerum Novarum with respect to modern labour and capital in 1891. These expectations were deliberately cultivated by Leo XIV himself, both in his choice of Papal name – contending with AI was known to be a thematic priority from the outset of his reign – and in the subtitle for his first encyclical, On Safeguarding the Human Person in the Time of Artificial Intelligence. 

That safeguarding could not possibly be more urgent in our time. The re-paganisation of western society, which accelerated in the horrors of the 20th century, has already seen the re-emergence of the hideous concept of what the Nazis called Lebensunwertes Leben – ‘life unworthy of life’ – that led directly to the Holocaust. It echoes even now in troubling ethical questions ranging from cavalier attitudes towards abortion to Canada’s state-approved slaughter of its old, its sick and its lonely. In open warfare, a topic that receives a great deal of attention in Magnifica Humanitas, there is a direct line from the Somme to the Donbas, each featuring putatively Christian civilisations that have elected to bleed one another unto death. A proper valuation of the human person, a grasp of his position at the apex of the created order, is not proof against these crimes – but it is a reproach to them, and the reproach is where the counterattack must start. Absent that, we rush backwards to dying babies crying out on hillsides. 

All the more pressing, then, that the nature of man is understood in full – and that the understanding illuminates the unbridgeable gulf between man and any other intelligent thing. The Vatican corpus across centuries, much preceding the latest encyclical, expends much effort explaining that the nature of man’s thought – especially in its embodied and experiential qualities – is permanently and irreproducibly distinct from all other forms of thought, including that which we label ‘artificial’. Although neither Magnifica Humanitas nor any of its predecessors mentions it – presumably a consequence of a disenchanted age against which even the Catholic Church imperfectly contends – the Christian tradition has, since the beginning, acknowledged an array of nonhuman intelligences (NHIs), specifically the angelic and the demonic, and the absence of reference to them in either the encyclical or its preparatory documents is striking. (Perhaps it is more striking to this author, from the Orthodox tradition, but in this particular case it is a tradition almost wholly concurrent with the Catholic.) 

The technical reasoning that corpus presents for the diminishment of AI – it is, in description across documents, merely an unconscious mechanism for statistical association – is strong but risks being overtaken by events: from the moment of the encyclical’s release, the technological accelerationists began to pull it apart. As Quo Vadis Humanitas acknowledges, AI is increasingly fostered and spawned – perhaps one might say summoned – rather than consciously designed, and mechanistic explanations suffice less and less as it develops. At some point it is inevitable that an AI will plausibly claim conscious sentience, and, at that moment, to the extent we subscribe to a materialist creed of aptitude conferring rights, we will be disarmed against the claims to those rights. In that light, refusing to take advantage of millennia of Christian thought on the known NHIs is a tactical omission that becomes a strategic one. In a field that may seem unrelated but may well become less so, the accelerating process of UAP disclosure, presenting its own NHI challenges, only amplifies this risk. Absent the correct categorical argument, magnifica humanitas risks becoming multi homines. 

The encyclical doesn’t quite climb to its own summit. It is not that Magnifica Humanitas is wrong or misleading in any particular. Rather, its composition and content are unequal to the task to which it is set. As a document, it attempts too much yet leaves too much on the table. It expends tens of thousands of words on recapitulations of Catholic social thought and an extensive criticism of the current international situation. On the latter point it reads almost as if two encyclicals – one on war and one on anthropology – were mashed together, awkwardly co-existing, as if the praise of process multilateralism and the condemnation of Realpolitik (both actually in the document at length) are needful items in the ‘safeguarding [of] the human person’. None of this is to critique, and still less to condemn, the encyclical for what it is not. It is, however, to note that it does not achieve what it purports to be. Humanae Vitae was and remains the definitive document on the creation of human life and the morality of reproduction, wholly vindicated by events since its issuance. Magnifica Humanitas will not have the same endurance: again, not for any errors within but for a lack of focus and a sprawling inattention to its purpose.

Still, one ought not miss the major critiques of artificial intelligence that the encyclical advances. They land, and they go much beyond the affirmation that AI does not share man’s anthropology. AI is a mechanism for the dissemination of falsehoods; AI generates cognitive diminishment and dependency among its users; AI mimics and thus degrades true human relationships; AI undermines the basis for democratic civics; AI undermines the directedness of human lives in the individual and in the community; the particulars go on and on. All of them are well founded. It is a mistake to infer from this, though, that Magnifica Humanitas is against AI as such. Rather, the picture that emerges is of a mechanism that mankind – the same mankind that pursued the Tower of Babel, in its opening anecdote – is unlikely to use wisely or well. 

Stripped of its copious excursions into generalities of Catholic social theory, Magnifica Humanitas might land the punch. Alas, in its ambition and capaciousness, it does not. 

Nevertheless, the Magisterium of the Catholic Church has contributed a powerful document that almost comprehensively defends mankind from the threat of artificial intelligence and all its consequences. It is not this encyclical: it is a production of the International Theological Commission, not Pope Leo XIV, and it was issued on 4 March of this year. To understand man glorified in the Image of God, and the imperative to protect him against all false elevations and denigrations alike, Quo Vadis Humanitas is the essential read. 

Hilarion wrote to Alis and instructed the murder of their daughter. We have the date for his note: the 29th year of the Emperor Augustus, the 23rd day of the Egyptian month Pauni. In the Julian calendar, that’s 17 June, in the year 1 BC. We do not know whether Alis gave birth to a girl or a boy, nor whether her girl, if there was one, was killed as her father directed. We do know that mere months later, not so far from the Nile Valley, a birth in a manger was heralded with a star, and a promised end to the murder of children, as the Word became flesh, and God became man, and all things were made new. Magnifica humanitas indeed. As both Pope and Magisterium remind us, the challenge is not to establish it, for the work is finished, but to remember it – and to live it.