Artificial Intelligence: The Pope Calls for Disarming AI

The relationship between the Pope and artificial intelligence has moved from offhand remarks to an official Church document. On 25 May 2026, Leo XIV published Magnifica Humanitas, the first encyclical of his pontificate — and the first in history entirely devoted to the age of artificial intelligence. The message is direct: technology exists to serve people, never to dominate them.

TL;DR

• Magnifica Humanitas was signed on 15 May and published on the morning of Monday, 25 May 2026; it is the first encyclical of Leo XIV.
• The Pope personally presented it in the Synod Hall at the Vatican — unprecedented for the public presentation of a magisterial document.
• The central theme is the call to “disarm AI”: preventing technology from becoming an instrument of domination, exclusion, or death.
• There is harsh criticism of the military use of AI and the concentration of technological power in a few private hands.
• For businesses, the document does not prohibit anything: it offers an ethical yardstick — human oversight, transparency, and the common good — for those adopting artificial intelligence.

This is not the first time a religious leader has commented on technology, but it is the first time artificial intelligence has received a document of this weight. Below, I have separated what the text actually says, without hype, and what it implies for those who work with technology every day — including us, who build AI-based products.

What is the encyclical Magnifica Humanitas

An encyclical is a solemn letter from the Pope that defines the Church's position on a topic. Magnifica Humanitas is a hefty tome: over 200 pages, divided into five chapters, dealing with the custody of the human person in the age of artificial intelligence.

The themes range from the wisdom of the Word and dialogue with the human sciences to Social Doctrine as community discernment, the evolution of Social Magisterium since Leo XIII, and the relationship between technology and dominion. It is not a technical manual or a list of prohibitions — it is a reflection on what it means to remain human when machines begin to influence decisions, shape behaviours, and even change how wars are conducted.

A symbolic detail drew attention: the Pope did not delegate the presentation. He personally went up to the Synod Hall to deliver the document to the public — unprecedented, as encyclicals are usually released by cardinals at press conferences. The gesture underscores the importance Leo XIV attributes to the topic.

A methodological note: an encyclical is neither prophecy nor technological guesswork. It is a doctrinal text, built with consultants and revised over years. When a document of this calibre chooses artificial intelligence as its subject, the signal is not about fashion — it is about permanence. The Church is saying that AI is no longer a passing trend but a structural issue for civilisation.

Why a Pope named Leo speaks again of revolution

The name chosen by the pontiff is no coincidence, and the encyclical itself makes the connection. In 1891, Leo XIII published Rerum Novarum, the document that inaugurated the modern Social Doctrine of the Church in the midst of the Industrial Revolution — addressing wages, working conditions, and the dignity of workers in the face of steam engines.

More than a century later, Leo XIV positions himself before another rupture: artificial intelligence. Magnifica Humanitas explicitly assumes this continuity by discussing the evolution of Social Magisterium since Leo XIII. The implicit message is strong: just as 19th-century industry required new rules to protect the worker, 21st-century artificial intelligence requires new rules to protect the person.

For those who work with technology, it is worth remembering the parallel. Every productive revolution concentrates power before distributing benefits. It was so with steam, electricity, and the internet. The question the Church raises is who remains in control — and at whose service.

“Disarming AI”: the heart of the document

The phrase that sums up the encyclical is “disarm AI”. At first glance, it sounds like a rejection of technology, but the Pope himself anticipates the misunderstanding. In the presentation, Leo XIV was categorical: “Disarming does not mean refusing technology, but preventing it from dominating humanity.”

Disarming, in the document's vocabulary, means freeing artificial intelligence from the logics that turn it into an instrument of domination, exclusion, or death, to reposition it at the service of the common good. It is not about erasing AI — it is about stripping it of the role of master and returning it to the role of tool.

This framing is important because it disarms (pardon the pun) the lazy reading that “the Pope is against AI.” He is not. The target is not technology itself, but the use that concentrates power and discards people. Those who create digital products should read this less as a sermon and more as a requirements specification: artificial intelligence needs brakes that are designed, not improvised.

Artificial intelligence and war: no algorithm makes conflict acceptable

The toughest point of the text is military use. The Pope denounces weapons based on artificial intelligence and sums up the position in a phrase already circulating worldwide: “No algorithm can make war morally acceptable.”

The argument is concrete, not merely moral. Autonomous systems make conflicts faster and more impersonal, and this lowers the threshold for the use of violence — when killing becomes a decision delegated to a model, the psychological and ethical barrier that holds the trigger disappears. The distance between the decision-maker and the dead increases, and responsibility dissolves into code.

This is the kind of debate that seems distant from a small business, but it is not. The same logic of “delegating critical decisions to the algorithm without supervision” appears, on a smaller scale, in credit systems, resume screening, and automated moderation. The principle is transferable: a decision that affects someone's life requires a responsible human in the loop.

The human cost behind artificial intelligence

The encyclical also looks beneath the shiny surface of AI. Two alerts stand out.

The first is the concentration of power. Leo XIV criticises the fact that the most influential technology of the decade is in the hands of very few private actors, and argues that the common good “cannot be left to the control of a few.” The decision on how artificial intelligence shapes work, information, and war cannot be a corporate backroom choice.

The second is the material human cost. The document records the precarious working conditions in labelling and training AI models — the invisible army that teaches the machines — and the exploitation, including of children, in the extraction of rare earths used in manufacturing equipment. The cloud, the text reminds us, has a factory floor and a mine.

Hence the phrase that perhaps best sums up the overall tone: “We have an urgent duty to remain deeply human.” Artificial intelligence, in the Pope's diagnosis, is not just a technical innovation — it is a force capable of redefining the very meaning of the human.

This material angle is often left out of discussions about AI, which focus on models and benchmarks. The encyclical forces the focus back to the chain: who labels the data, who extracts the minerals, who loses their job, and who profits. It is an uncomfortable reminder that every response generated by a chatbot has a human and environmental cost upstream.

What artificial intelligence changes for businesses

Here is where the document moves from the theological plane to the daily life of those who use technology. The encyclical does not regulate anyone — regulators are states and blocs like the European Union — but it offers a yardstick of principles that aligns well with the governance requirements already arriving.

In practice, the document's appeals can be translated into concrete product and operational decisions:

Principle of the encyclical	Practical translation for the business
AI at the service of the person, not domination	Maintain human oversight in sensitive decisions (credit, HR, customer service)
Rejection of opacity	Document where and how artificial intelligence is used and be able to explain results
Common good above the profit of a few	Assess impact on customers and workers before automating
Dignity of work	Use AI to augment the team, not just to cut it
Clear responsibility	Define a human responsible for each automated system

Interestingly, this ethical yardstick aligns almost point by point with regulation already in force. The EU AI Act, for example, classifies systems by risk level and requires precisely human oversight, transparency, and traceability in sensitive cases. In other words: what the encyclical asks for by moral conviction, the law is beginning to require by obligation. Companies that anticipate these principles will be ready when enforcement tightens.

None of these items stifle innovation. They merely avoid the dangerous shortcut of treating artificial intelligence as a black box that decides alone and no one audits. Those already following the advance of AI agents in businesses know that this question — “who answers when the agent makes a mistake?” — has ceased to be philosophical and become operational.

How to adopt human-centred artificial intelligence

At Agathas Web, we build products that rely on artificial intelligence every day — from service agents to marketing automations. Reading Magnifica Humanitas does not change the technology we use, but it reinforces a checklist we already advocated. It applies to any company adopting AI now:

1. Start with the problem, not the tool. AI is not a goal; it is a means. If automation does not improve the life of a customer or team, it should not exist.
2. Keep a human in the loop. Every decision with real impact (denying credit, firing, blocking a user) needs human review. The algorithm suggests; the person decides.
3. Be transparent. Tell the customer when they are talking to an artificial intelligence system. Trust does not survive deception.
4. Audit for bias. Models inherit the prejudices of their data. Test results by group before deploying.
5. Document. Record where artificial intelligence operates, with what data, and for what purpose. This is both ethics and legal protection.
6. Treat productivity gains as redistribution, not just cuts. AI frees up time — use that time for higher-value tasks, not just to reduce headcount.

Those who want to understand the technological landscape that makes this debate urgent can check our overview of Google I/O 2026 and what changes for Brazilian businesses. The speed of advancement is precisely what gives weight to the Pope's warning.

Conclusion: ethics is not a brake, it is a direction

It is tempting to read a religious document on artificial intelligence as a plea for conservative caution. But Magnifica Humanitas does not ask to brake — it asks to steer. The difference is enormous. Braking is stopping; steering is choosing where to go.

The Pope is not competing with engineers nor dictating software architecture. He is asking the question that technological euphoria usually skips: at whose service? For those building with AI, this is the best discovery question there is.

If your company is adopting artificial intelligence now, it is worth treating the document's principles as a free governance framework — human oversight, transparency, and focus on the common good. It costs nothing and avoids the most expensive mistakes. Want to review how AI is being used in your business responsibly? Talk to us — it is the kind of conversation we prefer to have before the problem, not after.