Pope Leo XIV’s AI Warning: Keep Machines Under Human Control

Pope Leo XIV’s recent AI message is less a rejection of technology than a warning about who controls it and what happens when human responsibility is delegated to machines. In addresses to the International Catholic Legislators Network on August 21 and the International Institute of Administrative Sciences on August 28, he called for AI to remain a servant of the common good—guided by human intelligence, conscience and law rather than by speed, profit or opaque systems.

The two speeches apply the central principle of his encyclical Magnifica Humanitas: technological progress should be judged by whether it protects human dignity, justice, peace and genuine human agency. 16

His warning to public administrators: people must not become victims of their inventions

Speaking to members of the International Institute of Administrative Sciences, Leo cautioned that the rapid development of AI raises the possibility that humanity could “become victim” of the inventions it has created. He specifically warned against entrusting decisions concerning human life to machines.

His response was not to abandon digital technology, but to humanize it. He urged public-administration researchers to find ways for digital systems to serve life and peace, keeping people—not automated processes—at the center of public decision-making.

That places a clear limit on automation: AI may assist administration, but it should not erase the human judgment, responsibility and moral accountability required when decisions affect people’s rights, welfare or dignity.

His warning to lawmakers: AI power is becoming concentrated

In his August 21 address to the International Catholic Legislators Network, Leo focused on the political responsibility of lawmakers. The Vatican described the meeting’s theme as “Human Dignity and Artificial Intelligence: Challenges, Chances and Control,” framing AI governance as one of the defining questions of the age.

The concern is not simply that some AI systems may make mistakes. It is that advanced technology and the power to shape its use may become concentrated among a small number of major economic and technological actors. Leo warned that this concentration could leave poorer countries increasingly dependent on wealthier ones and widen existing inequalities.

He also cautioned that innovation could become a vehicle for “ideological or economic colonialism”—a concern that turns AI governance into an international question, not only a domestic regulatory debate.

Algorithms can create a subtle form of domination

Leo described another risk: algorithms and platforms can exercise power without appearing to govern directly. In his account, algorithms may determine who is seen and who remains invisible, while digital platforms shape public discourse without adequate accountability. Automated systems can also subordinate workers’ dignity to optimization.

This is broader than the problem of technical bias. It concerns who sets the rules of visibility, participation and opportunity—and whether affected people have a meaningful way to challenge those decisions. When private systems influence public conversation or access to work without transparency, social power can move away from democratic institutions and public scrutiny.

Why he emphasized the family

Leo called the family the “first school of humanity” and urged legislators to prevent AI from eroding its role.

The point is that human formation begins in relationships that teach trust, responsibility, generosity, dialogue and concern for others. Those qualities are not equivalent to the speed or computational capacity of an AI system. Magnifica Humanitas likewise warns against confusing machine performance with human intelligence: AI systems can imitate certain functions of intelligence and outperform people in speed, but they do not possess human experience or embodiment. 16

For Leo, protecting family life is therefore part of AI policy. The question is not only whether a system is efficient, but whether its use supports the relationships and institutions through which people learn freedom, responsibility and solidarity.

The safeguards he urged lawmakers to build

Leo’s recommendations point toward a human-centered framework for AI governance:

• Protect fundamental rights and freedoms. Legislation should ensure that technological progress does not come at the expense of human dignity.
• Preserve privacy. Users should be protected from exploitation and inappropriate uses of personal information.
• Require transparency. People and institutions need to understand how emerging technologies are used and how consequential decisions are made.
• Strengthen democratic institutions. Digital systems should reinforce, rather than bypass, accountable public governance.
• Keep responsibility human. Decisions affecting human life and dignity should not be handed over to machines without meaningful human control.
• Prevent domination and excessive concentration. Governance should address the risks created when advanced AI is controlled by a small group of powerful actors.

These measures reflect the encyclical’s broader argument that technology is not socially neutral. It can foster participation and justice, but it can also intensify inequality, control and exclusion.

How the speeches extend Magnifica Humanitas

Published in May 2026, Magnifica Humanitas is explicitly devoted to safeguarding the human person in the age of AI. Its central test is whether technology remains oriented toward humanity rather than becoming an instrument of domination, exclusion or war. The encyclical calls for AI to be “disarmed” of those destructive logics. 51016

The August addresses translate that principle into responsibilities for two groups. Public administrators must ensure that digital systems serve citizens and preserve human judgment. Legislators must create rules that protect rights, privacy, transparency and democratic life. Researchers, meanwhile, have a role in identifying practical ways to make digital technologies serve life and peace.

Taken together, Leo’s warning is preventive. He is asking societies to establish human-centered rules while AI systems are still being developed and deployed—not after opaque platforms, concentrated power and automated decision-making have become too entrenched to challenge. The standard he proposes is straightforward: AI should expand the common good without displacing the human beings responsible for defining and defending it.