Antiqua et nova

Why this text matters

Before Antiqua et nova, the magisterial Catholic record on AI consisted of a small set of Francis’s addresses and messages — the 2024 World Day of Peace message, the World Communications Day message, the G7 address — plus the 2020 Rome Call for AI Ethics. These were powerful but dispersed. Antiqua et nova consolidates the work into a single doctrinal document with the joint authority of two Vatican dicasteries, in a format that closely resembles the older Holy Office notifications and is intended to function as such.

For the present site, the document matters because it is what the Catholic critique of the technocratic paradigm sounds like when it is applied directly and at length to the most consequential technology of the present moment. The diagnostic framework is recognizable — it descends from Guardini through Benedict XVI and into Laudato si’ — but the application is new. Antiqua et nova is the answer to the question: what does Catholic social teaching actually say about AI, in detail, when it sits down to write?

The argument in one paragraph

AI is real, novel, and the product of genuine human creativity — and therefore not to be despised — but it is not a moral subject and cannot become one. Human intelligence is embodied, relational, and truth-seeking in ways that artificial systems are not and cannot be, however impressive their statistical performance becomes. From this anthropology follows a developed ethic: AI deployments are evaluated by whether they honor or evacuate the embodied, relational, truth-seeking character of the human person. Specific concerns the document names: concentration of AI capability in a small number of corporations whose accountability is structurally weak; the rise of surveillance and manipulation enabled by predictive systems; the dehumanization of healthcare when AI displaces rather than supports the human carer; the development of lethal autonomous weapons; the formation of consciences by algorithmic environments; the temptation to treat artificial systems as objects of trust and reverence properly owed only to God (the document is unembarrassed about calling this idolatry). The constructive response is not refusal but discernment: human oversight at every consequential point, AI in subordinate roles to human moral judgment, protection of the most vulnerable, transparency, accountability, and the formation of consciences capable of resisting the technocratic paradigm. The slogan of the corpus, retrieved from Guardini, is the warning that we may become rich in technology and poor in humanity.

Key concepts

Embodied, relational, truth-seeking intelligence. The Catholic anthropology that anchors the whole document. Human intelligence is not reducible to information processing; it is the intelligence of an embodied creature in actual relations with other persons, seeking truth as a real condition of being. AI systems can produce useful outputs but do not have this kind of intelligence and cannot acquire it through capability scaling.

AI is not a moral subject. A precise doctrinal claim. AI systems do not have intentions, virtues, or accountability in the moral sense. Responsibility for AI outputs rests with the humans — designers, deployers, regulators, users — who make the system act. The widespread loose talk of “AI making decisions” obscures the actual chain of responsibility.

The Good Samaritan structure of care. The document repeatedly returns to the parable as a paradigm for what care of the vulnerable looks like, and argues that AI in healthcare and social services must support, not displace, that structure. The carer’s embodied presence, attention, and judgment are not optional efficiency overheads; they are constitutive of care.

Concentration of corporate AI power. Explicit warning about the structural risk of frontier AI capability sitting in a small number of corporations whose internal accountability is weak and whose public accountability is contested. The document does not name specific companies but the analysis is unmistakable.

Lethal autonomous weapons systems (LAWS). Specifically named as raising serious moral concern. The document continues the Vatican’s longer pattern of restrictive judgment on autonomous weapons; it does not declare them categorically impermissible, but the criteria it sets are stringent.

Idolatry. Used directly. The temptation to relate to AI systems as objects of trust, reverence, or final authority properly owed only to God is named as idolatry. The vocabulary is theological, not metaphorical.

The technocratic paradigm. Carried over from Laudato si’. AI is treated as the current most acute instance of the paradigm. The same diagnostic frame applies; the specific implementation is new.

The wisdom of the heart. From Francis’s 2024 Communications Day message. The conviction that genuine human knowledge integrates intellect, emotion, embodiment, and relation in a way AI cannot, and that this integration is what Christian formation cultivates.

Where it sits on the map

On the preserve-limits ↔ accelerate-transformation axis, Antiqua et nova sits modestly to the limits side — like Laudato si’, it is not against acceleration per se but is against the technocratic paradigm that treats acceleration as self-justifying.

On the two independent concerns axis, the document is high on idolatry concern (extremely high, given the explicit naming of AI idolatry as a category) and moderate on technology-as-central-to-Christian-hope (it honors AI as a product of human creativity, refuses to make it central to hope in the strong sense).

Pair with Laudato si’ for the framework and with the ITC’s Quo vadis, humanitas? for the theological extension into transhumanism and digital religion.

Best passage to verify

The document is numbered throughout (a Vatican convention), which makes citation unambiguous. The most-cited sections will likely be:

• The opening section on the title’s “ancient and new” framing — what is genuinely new in AI and what is an ancient pattern under new technical conditions.
• The anthropology section developing embodied, relational, truth-seeking intelligence.
• The section on healthcare and the Good Samaritan structure of care.
• The section on lethal autonomous weapons.
• The explicit naming of AI idolatry.
• The closing section on the formation of conscience inside algorithmic environments.

A verified pull-quote from the Vatican English text should be inserted here before final publication. The Vatican publishes the document in all major languages on vatican.va.

What it gets right

Three things Antiqua et nova did that the broader Christian conversation needed.

First, it gave the magisterial position on AI a single articulated document, instead of dispersing it across addresses and messages. Before 2025, anyone trying to summarize “what the Catholic Church thinks about AI” had to assemble fragments from World Day messages, the Rome Call signatories, and occasional dicastery statements. The note replaces that fragmentary record with a single doctrinal text that can be cited as a unit.

Second, it made the anthropology load-bearing. The document’s whole ethical analysis follows from the prior commitment to embodied, relational, truth-seeking intelligence. This is exactly the right order — the ethics is downstream from the anthropology, not bolted on as policy preference. Catholic theology of technology done well always works this way; the document is a good example of the form.

Third, it named idolatry directly. Most secular AI-ethics discourse cannot name the religious dimension of the problem because its vocabulary does not allow for it. The Vatican’s vocabulary allows for it, and Antiqua et nova uses the word. This is a significant rhetorical and analytic move. The contemporary temptation to relate to AI systems as authoritative arbiters of truth, taste, or value is more accurately described as idolatry than as any of the secular alternatives, and naming it as such makes resistance theologically intelligible.

What to argue with / what it misses

The most serious criticisms are coming from inside the Catholic conversation rather than outside it.

First, the constructive theology of AI engineering is gestured at, not developed. The document honors AI as a product of God-given creativity and acknowledges its real goods (especially in medicine), but it does not work out, with the depth that Dessauer’s framework would allow, what the AI engineer’s vocation actually consists of when done well. The Catholic conversation needs the constructive Build-and-Heal theology of the AI engineer in addition to the critique; Antiqua et nova does not supply it.

Second, the eschatological dimension is muted. The document is strong on anthropology and ethics; it is comparatively cautious on the explicit eschatological questions about life extension, mind-upload fantasies, transhumanism, and digital immortality. These are the questions where Quinzio’s framework and the contemporary tech-apocalyptic discourse intersect with magisterial teaching, and where readers expecting Antiqua et nova to do that work will find it doing somewhat less. The ITC’s Quo vadis, humanitas? (2026) was published shortly after, partly to fill this gap.

Third, the political-economic specificity could be sharper. The note warns about concentration of corporate AI power, but does not — at the dicastery’s level of generality — name specific firms, specific business models, or specific governance arrangements. This is a fair criticism but also a generic feature of doctrinal documents; the application to specific contexts is left to bishops, scholars, and laypeople.

Fourth, secular AI-ethics critics sometimes find the document either too theological (when read by secular liberalism) or too cautious about specific technical interventions (when read by AI-safety maximalists). Both readings are partial. The document is exactly as theological as the analysis requires, and exactly as cautious about specific technical interventions as a doctrinal note should be when the specific interventions are evolving rapidly.

Later influence

The document is recent (2025) and its full institutional uptake is still developing. The immediate downstream documents:

• The ITC’s Quo vadis, humanitas? (2026) extends the eschatological and transhumanist analysis that Antiqua et nova left under-developed.
• A growing body of bishops’ conference statements applying the framework to local contexts (notably in Europe and Latin America).
• An emerging Catholic AI-ethics literature led by Paolo Benanti, Paul Scherz, and others, taking the document as the framework against which more specific analysis is built.

Outside Catholic circles, the document has been cited respectfully by parts of the secular AI-ethics community, by some technology journalists, and by signatories of the original Rome Call. Its long-term institutional influence will depend on how seriously Catholic dioceses, universities, and the broader Catholic-affiliated intellectual world treat it as a working framework.

How it speaks to AI, platforms, and modern work

This is the AI document. Nearly every section speaks directly to the contemporary deployment.

On large-language-model deployment. The anthropology of embodied, relational, truth-seeking intelligence is the explicit corrective to the temptation to treat LLM outputs as authoritative knowledge. The model can produce plausible text; it cannot stand in for the embodied human carer, teacher, judge, journalist, or pastoral worker whose role is partly constituted by being present. Deployments that displace those roles without preserving their embodied character fail the Antiqua et nova test.

On AI medicine. The Good Samaritan structure of care is the explicit standard. AI is welcome where it augments the human carer’s attention, judgment, and presence; it fails the test where it displaces them. Most current medical AI is somewhere on the line; the document gives criteria for distinguishing.

On agentic AI. Less explicitly addressed, but the “AI is not a moral subject” principle has direct implications. An “agentic” AI that takes consequential actions cannot be the moral responsible party; that responsibility remains with the humans who deployed it. The corporate vocabulary that talks about AI agents as if they were responsible actors obscures the actual chain of accountability and should be resisted.

On lethal autonomous weapons. The document continues the Vatican’s restrictive line. The criterion is human moral judgment at the consequential moment, which LAWS by definition remove.

On AI as quasi-religious mediator. The idolatry framing applies most sharply where AI systems are positioned as sources of advice, judgment, or comfort that bypass the embodied human relationships through which God’s grace ordinarily reaches the person. The contemporary phenomena of AI “spiritual advisors,” AI therapy chatbots, and AI-mediated bereavement assistance are all under this analysis.

On the corporate form. The warning about concentration of AI capability in a small number of corporations should be read alongside the more general critique of the technocratic paradigm in Laudato si’. The document does not propose a specific corporate-governance solution; the analysis is left as a standing problem that bishops, scholars, and laypeople are expected to take up.

Read next

• Laudato si’ — the framework document; Antiqua et nova applies its diagnosis to AI.
• Quo vadis, humanitas? (ITC, 2026) — extends the analysis into transhumanism, digital religion, and the future of humanity.
• Letters from Lake Como — the deep ancestor of the technocratic-paradigm framework.
• The Technological Society — the Protestant sociology that the Catholic anthropology now sits beside.
• Quinzio — the eschatological pressure point that Antiqua et nova does not fully engage with but cannot avoid.

Source note

Antiqua et nova is published in all major languages on vatican.va. The English text is the working English for most readers. As with all dicastery documents, contested passages should be read against the Italian and Latin originals. Numbered paragraphs make citation unambiguous.

The companion documents in the Vatican AI corpus — the 2020 Rome Call for AI Ethics, Francis’s 2024 World Day of Peace and Communications Day messages, the G7 address, and the ITC’s Quo vadis, humanitas? (2026) — should be read together with Antiqua et nova for the full institutional picture. Treating the document in isolation underestimates its place in a developing corpus.

This commentary draws on the Vatican English text, on Paolo Benanti’s contemporary work on AI ethics, on the post-publication theological reception in La Civiltà Cattolica and the Journal of Moral Theology, and on the Francis (and the Vatican AI corpus) thinker page for the institutional context.