Leo XIV at Geneva: A spiritual diagnosis of AI's category error

On July 8, 2026, Pope Leo XIV sent a letter through Vatican Secretary of State Cardinal Pietro Parolin to the AI for Good Global Summit in Geneva, addressing the UN-affiliated body on a specific danger: AI systems that treat the human person as a data profile, optimizing behavior in ways that bypass genuine interiority and relationship. The letter explained that his encyclical Magnifica Humanitas, presented at the Vatican on May 26, 2026, and born from conversations with scientists, political leaders, and parents, was also driven by 'troubling accounts of the potential misuses of algorithms and by the loss of human agency in critical areas.' The Geneva address applies that encyclical's anthropology directly to the summit's work.

Benedict XVI, in his Wednesday Audiences on Caritas in Veritate, articulated the premise underlying Leo XIV's intervention: the Church does not offer technical solutions to social problems, but does insist on the anthropological premises without which no technical solution can be oriented toward genuine human development.[^1] Unlimited confidence in technology 'ultimately shows itself to be illusory' — not because technology is suspect, but because the human person is the condition of its proper use, not merely its product.

The technocratic paradigm as structural disorder

The Dicastery for the Doctrine of the Faith's 2025 document Antiqua et Nova names the specific error Leo XIV's Geneva letter targets: what Pope Francis had called the 'technocratic paradigm,' the assumption that 'reality, goodness, and truth automatically flow from technological and economic power as such.'[^2] In this frame, AI becomes dangerous precisely when it is assumed to be self-sufficient — when its pattern-matching outputs are treated as adequate accounts of human motivation, decision, or identity. The document argues that AI should instead serve 'the common good of the entire human family,' understood as the social conditions allowing people to reach their fulfillment.[^2] The problem is structural: an AI system built to optimize a measurable proxy for human behavior will consistently mistake the proxy for the person.

Jacques Maritain's distinction between the individual and the person gives this critique philosophical precision.[^3] The individual, in Maritain's account, is a part of the social whole, subject to its ordering and accountable to its structures. The person, ordered to a transcendent end, exceeds any social totality, including any data architecture. These two accounts of the human being differ in kind, not merely in degree. To reduce a human being to a behavioral profile is to operate entirely within the register of individuality while remaining blind to personhood. Leo XIV's Geneva letter insists on the distinction: persons carry a depth of interiority no dataset can exhaust, because their final orientation is toward an end that no optimization target can represent.

What the intellect and will actually do

Aquinas, in the Summa Theologiae, grounds this claim in the structure of the soul itself. As Ferdinand Ulrich notes in Homo Abyssus, drawing on Aquinas's account: 'to know what you may will, or what you may understand, does not belong to the perfection of my intellect; but only to know the truth in reality.'[^4] The intellect's proper object is truth as such, and the will's proper object is the good as such, neither of which is a measurable output. A pattern-matching system trained on behavioral data can predict what a person has done; it cannot orient itself toward what a person ought to become. That gap is the anthropological fact Leo XIV's Geneva letter asks policymakers and technologists to reckon with.

Craig Steven Titus, Paul Vitz, and William Nordling, in the Catholic Christian Meta-Model of the Person, argue that the unity of body and soul in the human person means personal identity cannot be disaggregated into measurable components without remainder.[^5] The person is a 'personal unity and individual substance, with unique human, moral, and spiritual goals or ends,' a description that applies to the whole, not to any subset of behavioral or neurological data. Any algorithmic model that treats those data subsets as the person has already committed the error the encyclical names.

A witness without a lobby

The pope's address to a UN-affiliated summit continues a long pattern: the Church engages international institutions as a witness to an anthropology those institutions otherwise lack the vocabulary to articulate. Leo XIV's letter to Geneva is not a request for a specific regulatory outcome, though it may inform one. Its primary function is diagnostic, identifying in the present AI moment the same Promethean error that Catholic social thought has named since Rerum Novarum: the substitution of a technical system for the irreducible interiority of the person.

That diagnosis carries a specific pastoral consequence for mental health and formation. When AI tools enter therapeutic, educational, or pastoral contexts, they bring the same structural temptation: to treat algorithmic summaries of a person's history, symptom profile, or behavioral tendencies as adequate to the person themselves. The Catholic Christian tradition's insistence on the unity of body and soul, on the intellect's orientation toward truth and the will's orientation toward good, supplies the anthropological commitments that make accompaniment possible, the recognition that no profile, however detailed, exhausts the person seated across from you. That recognition is the condition without which accompaniment collapses into case management.

The disorder Leo XIV names in Geneva is old: it is the Promethean temptation to replace what cannot be engineered with what can be measured. The tradition has always known the difference. The question is whether the institutions now governing AI will learn it before the cost of the confusion becomes irreversible.

References

[^1]: Pope Benedict XVI. (2009, July 7). General audience on Caritas in Veritate [Wednesday Audience transcript]. Vatican Publishing House. In WednesdayAudiences.pdf.

[^2]: Dicastery for the Doctrine of the Faith. (2025). Antiqua et Nova: Note on the relationship between artificial intelligence and human intelligence. Libreria Editrice Vaticana. §§54–55.

[^3]: [Lecture recording]. Segment 5: Maritain's Personalist Philosophy: Person, Individual, Common Good, and Critique of False Thomism (00:22:43–00:48:03). In Theological and philosophical foundations of justice and community in Christian life.

[^4]: Ulrich, F. (1998). Homo Abyssus: Das Wagnis der Seinsfrage. Johannes Verlag Einsiedeln. Introduction, citing Aquinas, Summa Theologiae I, q. 107, a. 2.

[^5]: Titus, C. S., Vitz, P. C., & Nordling, W. J. (2020). Personal wholeness (unity). In A Catholic Christian meta-model of the person (Ch. 8). Divine Mercy University Press.