Inside the Vatican's 2025 Internal AI Governance Guidelines

The decree arrived quietly. On December 16, 2024, the Pontifical Commission for Vatican City State signed Decree No. DCCII — a set of guidelines governing the use of artificial intelligence within the world's smallest sovereign state. Fifteen days later, on January 1, 2025, they came into force. No press conference. No viral announcement. Just the Church, as it has done for two millennia, doing what it believes is right, and leaving the world to catch up.

Most coverage of Vatican AI focuses on the philosophical: papal encyclicals, theological debates about consciousness, the question of whether a machine can have a soul. That conversation matters. But there's another story — a more operational, more immediate one — unfolding in the 44 hectares between the Tiber and St. Peter's Square. The Vatican has written actual rules. And they are more sophisticated than anyone expected.

What the Document Is — and Isn't

First, important context. The 2025 guidelines are not a papal encyclical. They are not magisterial teaching. They are an internal governance document — the equivalent of a company policy or a government agency's operating procedures — issued by the Governorate of Vatican City State, the administrative body that runs the day-to-day operations of the tiny city-state.

This distinction matters. The Governorate oversees roughly 800 employees: Vatican Museums staff, the Gendarmerie Corps, firefighters, postal workers, healthcare workers, archivists, groundskeepers for the Vatican Gardens, and the teams managing the Pontifical Villas at Castel Gandolfo. The guidelines apply to all of them — and to any external contractors, suppliers, or professionals who engage with AI while working within Vatican jurisdiction.

Think of it less as theology and more as policy — except policy written by an institution that has been thinking about human dignity and the common good for twenty centuries.

The Core Framework: Five Pillars

The guidelines are built around five interconnected principles, each rooted in Catholic social teaching but expressed in the language of modern AI governance.

1. Human Dignity as the Non-Negotiable Floor

Every provision in the document flows from a single premise: AI must serve the human person, not the other way around. The guidelines describe this as an "anthropocentric approach" — AI systems must be designed and deployed in ways that respect the unique dignity of every individual, protect human autonomy, and never reduce a person to a data point.

This is not rhetorical. The document specifies concrete prohibitions: AI may not be used to manipulate individuals through subliminal techniques, to exclude people with disabilities from services, or to make decisions that deny persons their fundamental rights. Human judgment must remain sovereign in high-stakes domains.

2. Transparency and Labeling

One of the most operationally specific provisions in the guidelines concerns AI-generated content. All content produced by artificial intelligence within Vatican jurisdiction must be clearly labeled — marked with the notation "IA" (the Italian abbreviation for intelligenza artificiale) — so that users can always distinguish between human and machine output.

This rule applies to texts, images, audio, and audiovisual content. In practice, it means that any Vatican communications — press releases, educational materials, museum guides, administrative documents — that incorporate AI-generated elements must disclose that fact. The principle is simple: you have a right to know when you are reading a machine.

3. Intellectual Property and Authorship

The Vatican's position on AI-generated intellectual property is unambiguous — and unusually clear compared to most institutions still wrestling with the question. The decree states that the Governorate of Vatican City State holds exclusive ownership of the authorship rights and economic rights over any content created through AI within its territory or jurisdiction.

This includes textual works, musical compositions, photographs, audiovisual content, radio broadcasts, and works of figurative art. If AI helps create it on Vatican soil, the Vatican owns it. The guidelines also explicitly prohibit using AI to violate the copyright of existing creative and artistic works — a provision aimed squarely at the generative AI systems that have trained on vast libraries of human-created work.

4. Cultural Heritage: A Special Case

The Vatican holds one of the most significant cultural heritage repositories on earth. The Vatican Museums. The Apostolic Library. The Secret Archives. The Sistine Chapel. Centuries of manuscripts, artworks, and artifacts that belong, in a meaningful sense, to all of humanity.

The guidelines treat this heritage with corresponding care. AI may be used to enhance preservation and accessibility — the Vatican has already partnered with Microsoft to digitize portions of its archives — but only under strict conditions. Any AI application in cultural heritage must align with the Governorate's institutional purposes, must not harm the physical integrity of the artifacts, and must comply with internationally recognized standards for restoration and conservation.

The message is clear: the machine serves the artifact, not the reverse.

5. Continuous Oversight: The AI Commission

Perhaps the most structurally significant element of the guidelines is the creation of a dedicated oversight body: a five-person AI Commission, presided over by the Secretary-General of the Governorate. This commission has a clear mandate — to prepare future laws and regulations, provide expert opinions on specific AI deployments, and continuously monitor AI activities within Vatican City State.

The guidelines describe this commission as a compass for future regulatory action. Its work is explicitly framed as ongoing: the decree acknowledges that AI is evolving rapidly, and that the guidelines themselves are a starting point, not a final word. New laws and specific regulations will follow as the Commission develops its understanding of the technology in context.

Why This Matters Beyond Vatican City

An 800-person city-state writing AI policy might seem like a footnote. It isn't. The Vatican's influence has never derived from its physical size. It derives from its moral authority, its institutional continuity, and its unmatched ability to speak to 1.3 billion Catholics — and beyond them, to anyone who takes questions of human dignity seriously.

The 2025 guidelines arrive in a global context where most governments are still arguing about first principles. The European Union's AI Act is the most ambitious secular framework, but it is enormously complex and still being implemented. The United States has no comprehensive federal AI legislation. China's approach prioritizes state interests. In this landscape, the Vatican's document stands out not for its regulatory teeth — it has limited enforcement power — but for its clarity of values.

The guidelines do something that most AI governance documents struggle with: they start from the human person and work outward. Not from technical capabilities. Not from economic efficiency. Not from national security. From a clear, theologically grounded answer to the question: What is a human being, and what does that mean for how we build our tools?

The Bigger Picture: Antiqua et Nova

The governance guidelines did not emerge in isolation. Two weeks after they took effect, on January 28, 2025, the Vatican released a sweeping philosophical document: Antiqua et Nova: Note on the Relationship Between Artificial Intelligence and Human Intelligence. Jointly issued by the Dicastery for the Doctrine of the Faith and the Dicastery for Culture and Education, and approved by Pope Francis on January 14, the 118-paragraph text provides the theological scaffolding for everything the governance decree operationalizes.

Antiqua et Nova draws on Scripture, Catholic teaching, and classical philosophy to argue that true intelligence — human intelligence — is not merely functional or data-driven. It is spiritual, embodied, relational, and oriented toward truth and love. AI, however sophisticated, cannot replicate this. It can simulate output; it cannot generate meaning. Its most memorable formulation: "Only the human person can be morally responsible. AI should be guided by human intelligence — not the other way around."

Together, the governance guidelines and Antiqua et Nova represent something genuinely new: a coherent Vatican doctrine on artificial intelligence, complete with philosophical foundations, ethical principles, and operational rules. It is the most integrated institutional response to AI that any religious body has produced.

What Comes Next

The Vatican has been explicit that the guidelines are a beginning. The AI Commission is charged with developing concrete laws and regulations within a year of the decree's enforcement — meaning specific rules should be forthcoming through 2025 and into 2026. These will address implementation details that the current guidelines leave open: enforcement mechanisms, technical standards for AI labeling, procurement criteria for AI systems used by Vatican entities, and protocols for AI use in healthcare within Vatican's medical facilities.

More broadly, the Vatican is positioning itself as an active participant in global AI governance conversations. Pope Leo XIV — who succeeded Francis in 2025 — addressed the Second Annual Rome Conference on AI, Ethics, and Corporate Governance in June 2025, emphasizing the Church's particular concern for young people's intellectual and neurological development in an age of generative AI. The Rome Call for AI Ethics, which the Vatican helped launch in 2020 with Microsoft and IBM, continues to expand its roster of signatories.

Two thousand years ago, the Church developed a sophisticated jurisprudence for navigating the moral complexities of its age. It is now, methodically and with characteristic patience, doing the same for the age of artificial intelligence. The decree is quiet. The implications are not.