Magnificent humanity: Leo XIV denounces AI at the service of the powerful

Context

Six weeks after the publication of Pope Leon XIV's encyclical Magnifica humanitas, Vatican Radio continues its series of analyses. The sixth episode, broadcast on July 12, 2026, in the German edition, focuses on one of the most striking passages of the text: artificial intelligence "is not neutral." It serves purposes; and the current trajectory of this technology, the Pope emphasizes, is not oriented for the benefit of the majority of humanity. This broadcast comes on the eve of major European deadlines and as the inter-dicasterial commission on AI, which met for the first time on June 17, prepares its first guidelines.

The Facts

Five points are highlighted. One: the alleged neutrality of algorithms is a fiction; every objective function reflects a moral choice. Two: training data selects an image of man, predominantly Western and Anglophone. Three: the economic concentration of the sector, where a handful of companies capture most of the world's value, creates civilizational dependence. Four: military and advertising uses structure development, not pastoral or educational needs. Five: the reader is invited to move from fascination to responsibility.

Doctrinal Analysis

Magnifica humanitas follows the line of Caritas in veritate (Benedict XVI, 2009), which reminded us that technology is never separable from morality and that its direction always hides a direction of the heart. Leon XIV radicalizes this intuition: what is at stake is anthropology. The creature in the image of God (Gn 1, 27) cannot delegate to the machine either moral decision, inner life, or political discernment. Saint Thomas reminds us, in the Summa Theologica, that practical reason belongs properly to the human person, who alone can judge and will in conscience. No system, no matter how sophisticated, can bear the moral act, which requires freedom and responsibility.

Stakes for the Church and the Faithful

Three stakes. First, discernment: the Vatican inter-dicasterial commission, whose first meeting took place on June 17, 2026, at the Palazzo San Calisto, must publish its first guidelines in the coming months. Then, education: European dioceses are invited to integrate critical training in AI into the catechesis of adolescents and in seminaries. Finally, resistance: refusing the blind use of large models to write homilies, catechesis, and teachings. Christ is not announced by algorithm.

Critical Reading and Blind Spots

Two blind spots. On the one hand, the encyclical remains discreet about the articulation between social doctrine and the emerging European law (AI Act, 2024). It is in Europe that a regulation is being played out to which the Church can concretely contribute. On the other hand, no mention is made of the energy crisis induced by generative AI: data centers consume as much as a medium-sized country. Sobriety, dear to Laudato si', awaits to be applied to data itself.

To Meditate and Act

Take up Benedict XVI's exhortation to the Bundestag in 2011: reason without the divine becomes deaf. Refuse to be "clients" of large models to become again persons who discern. Pray for developers and legislators. Join Catholic circles of digital ethics that, in France, attempt to build a Christian intelligibility of AI.