Pope Leo XIV’s first encyclical arrives with a title that sounds like it belongs to the age of algorithms. But as the Vatican’s message is read alongside the reporting and reactions it has already sparked, it becomes clear that the document is less a treatise on artificial intelligence than a diagnostic tool—one that uses AI as a mirror to reflect older, stubborn problems: concentrated power, the slow erosion of democratic accountability, and the emergence of a “tech elite” whose influence can outgrow the institutions meant to check it.
In other words, AI is not presented as the main character. It is the lens. The encyclical’s deeper subject is what happens when decision-making, persuasion, and access to information become increasingly mediated by systems that most people cannot audit, most governments cannot fully control, and many voters cannot meaningfully understand. The Vatican’s framing is notable not only because it comes from a religious authority, but because it lands in the middle of a policy debate that has often been dominated by technical jargon and corporate incentives. Pope Leo XIV’s approach is to translate the stakes into moral and civic language—without pretending that the technology itself is the root cause.
That translation matters. For years, public discussion about AI has tended to split into two camps: those who focus on capabilities and risks in narrow terms (bias, job displacement, misinformation), and those who focus on governance mechanisms (regulation, transparency, audits). The encyclical, at least as described in early coverage, tries to connect these threads to a more structural question: who gets to steer society when the steering wheel is increasingly embedded in software?
The encyclical’s argument begins with a familiar observation, but one that feels newly urgent in the AI era: power rarely stays evenly distributed. It concentrates—through economies of scale, through network effects, through data advantages, and through the ability to set standards. When AI systems are deployed at scale, they can amplify this concentration. Not because every model is inherently authoritarian, but because the infrastructure around them tends to be. Compute, proprietary datasets, cloud platforms, distribution channels, and the talent pipeline that builds and maintains these systems are not evenly spread across the world. The result is a shift in leverage: the entities that can deploy AI effectively gain influence over markets, politics, and culture, often faster than democratic processes can respond.
This is where the encyclical’s “concentrated power” theme becomes more than a generic warning. The Vatican’s concern is not simply that a few companies might become wealthy. It is that the locus of decision-making can move away from public scrutiny. When AI is used to allocate resources, prioritize content, score risk, or recommend actions, it can quietly reshape what counts as “choice.” People may still click buttons and cast votes, but the environment in which those choices are made—what options appear, what information is surfaced, what outcomes are predicted—can be engineered by actors who are not accountable to the public in the same way elected officials are.
The encyclical’s emphasis on democracy follows naturally. Democracy depends on visibility: citizens must be able to see how decisions are made, who benefits, and what trade-offs are being chosen. Accountability requires more than formal elections; it requires the ability to contest outcomes and demand explanations. AI systems, especially when they are integrated into government services, advertising ecosystems, or platforms that mediate public discourse, can erode that visibility. Even when an AI system is technically “working as intended,” the public may not know why a particular outcome occurred, or whether the system’s logic reflects values that align with democratic norms.
Early reactions to the encyclical have highlighted its insistence that the erosion of democracy is not always dramatic. It can be incremental. A society can lose democratic resilience without ever experiencing a single coup. If citizens gradually become dependent on opaque systems for information, if political messaging becomes increasingly personalized and difficult to trace, if oversight bodies struggle to keep pace with technical complexity, then democratic participation can become performative rather than substantive. The encyclical’s framing suggests that AI can accelerate this drift by making manipulation easier, targeting more precise, and accountability harder to enforce.
One of the most striking elements in the coverage is the encyclical’s attention to a “tech elite”—a phrase that captures a specific kind of power. This is not merely the power of wealth, nor only the power of innovation. It is the power that comes from being positioned at the intersection of data, infrastructure, and narrative control. The tech elite, in this view, is not defined by religion or nationality, but by function: those who can shape the world to their own advantage because they control the tools that increasingly mediate reality.
This is where the encyclical’s unique take becomes apparent. Rather than treating AI as a neutral instrument that can be regulated after the fact, the Vatican appears to argue that AI changes the structure of influence. When AI systems are embedded into everyday life—recommendation engines, automated decision systems, predictive analytics—power becomes less visible and more ambient. It does not always announce itself. It operates in the background, shaping what people see, what they are offered, and what they are told is likely to happen next.
The encyclical’s moral vocabulary—fairness, human dignity, responsibility—serves a practical purpose here. It pushes the conversation beyond “Is AI good or bad?” toward “What kind of society does AI tend to produce when deployed under current incentives?” That question is uncomfortable for everyone involved: for companies that want to move fast, for governments that want to adopt new tools without slowing down, and for citizens who want benefits without friction. But it is also the question that policy debates often avoid because it requires confronting trade-offs.
Consider the issue of transparency. In theory, transparency is a solution: disclose how systems work, publish documentation, allow audits. In practice, transparency can be limited by intellectual property claims, security concerns, and the sheer complexity of modern models. Even when documentation exists, it may not be understandable to non-experts. The encyclical’s democratic concern implies that transparency must be more than a checkbox. It must enable meaningful contestation. Citizens and oversight bodies need enough information to challenge outcomes, not just to satisfy curiosity.
This is also where the encyclical’s focus on concentrated power intersects with the “tech elite” theme. If only a small group can interpret, validate, and improve AI systems, then that group gains a gatekeeping role. They become the translators between raw data and real-world consequences. Over time, that translation power can become political power. It can influence which problems get solved, which risks get prioritized, and which communities bear the costs of experimentation.
The encyclical’s approach, as reflected in the reporting, suggests that the Vatican sees AI as a catalyst for a broader shift in governance. The shift is not necessarily toward overt authoritarianism. It can be toward technocratic rule-by-default, where decisions are justified by technical necessity rather than democratic deliberation. When AI outputs are treated as authoritative because they are “data-driven,” the burden of proof can shift. Instead of asking whether a decision is fair, the question becomes whether the system was trained correctly. Instead of debating values, the debate becomes about performance metrics.
That is why the encyclical’s emphasis on eroding democracy resonates beyond the church. Many democracies already struggle with polarization, misinformation, and institutional trust. AI can intensify these problems by scaling content production, automating persuasion, and personalizing narratives. But the encyclical’s deeper point is that even when AI is used responsibly, the mere presence of opaque decision systems can weaken democratic agency. People may feel that outcomes are predetermined by models they cannot interrogate. When that feeling spreads, democratic legitimacy suffers.
At the same time, the encyclical does not appear to argue for rejecting AI outright. The Vatican’s stance, as described, is more diagnostic than prohibitive. It treats AI as a lens that reveals patterns already present in society. That framing is important because it avoids the trap of scapegoating technology. It also avoids the opposite trap of techno-utopianism, where AI is treated as a cure-all for social problems. The encyclical’s implicit message is that the moral and civic questions remain regardless of the tool: Who holds power? Who is accountable? Who benefits? Who gets heard?
This is where the encyclical’s timing and audience matter. Pope Leo XIV’s message is not aimed only at engineers or regulators. It is aimed at civic leaders, faith communities, and the broader public—people who may not follow model architectures but who understand the lived experience of fairness and injustice. By using AI as a lens, the Vatican can speak to a generation that encounters technology daily while still grounding the conversation in enduring ethical concerns.
The Vatican’s engagement also signals something about the direction of global policy discourse. In recent years, AI ethics has often been framed as a competition between frameworks: corporate guidelines, academic principles, national strategies, and international declarations. Faith-based and civic leaders have sometimes been peripheral to these debates, despite their influence on public values and community organizing. The encyclical’s prominence suggests that the Vatican intends to be part of the conversation—not as a technical authority, but as a moral and civic one.
That moral authority can be influential in two ways. First, it can help translate complex issues into language that communities can act on. Second, it can pressure institutions to consider legitimacy, not just compliance. A regulation that forces disclosure but does not enable contestation may still leave democracy weakened. A governance model that focuses on safety testing but ignores who controls deployment may still concentrate power. The encyclical’s framing encourages readers to ask whether the system’s structure aligns with democratic ideals.
There is also a subtle but important point about “diagnosis.” The encyclical is not merely warning about future dangers; it is diagnosing present dynamics. Concentrated power, eroding democracy, and the rise of a tech elite are not new phenomena. What is new is the speed and scale at which AI can embed these dynamics into everyday