On Monday, two visions of our AI future collided, though they were separated by thousands of miles and centuries of institutional history. In Rome, Pope Leo XIV, leader of the world’s 1.3 billion Catholics, released Magnifica Humanitas, his first major encyclical. It was a sweeping, philosophical warning against the concentration of technological power. Simultaneously, in a move that could have been ripped from the encyclical’s subtext, OpenAI, the engine of the current AI boom, announced its latest strategic partnership, this time with Brazil’s largest media conglomerates, Grupo Folha and Grupo UOL.
One was a profound moral critique of the new technological world order. The other was a pragmatic, tactical expansion of that very order. Taken together, they offer a stark portrait of the central tension of the AI era: the chasm between our societal aspirations for this technology and the commercial realities of its deployment. While one side speaks of human dignity and the common good, the other is busy signing the deals that will define our new information reality. The question is whether these two conversations can ever truly meet.
A Warning from the Old World
A papal encyclical is not a research paper or a whitepaper. It is one ofthe most significant teaching documents a pope can issue, intended to guide faith and morals. For Pope Leo XIV to dedicate his very first one to artificial intelligence signals a profound engagement with what he sees as a hinge point for civilization. But to read Magnifica Humanitas (“Magnificent Humanity”) as a simple screed against algorithms is to miss the point entirely.
The 200-page document is less a technical analysis of large language models and more a sociological diagnosis of our time, using AI as its primary lens. The Pope’s central argument is that AI is not creating entirely new problems, but rather acting as a powerful accelerant for existing pathologies: wealth inequality, the erosion of democratic institutions, and the unchecked influence of a small, unaccountable elite.
“When such power is concentrated in the hands of a few,” he writes, “it tends to become opaque and evade public oversight, increasing the risk of distorted forms of development that give rise to new dependencies, exclusions, manipulations and inequalities.”
This is the core of the Vatican’s critique. It’s not about killer robots or esoteric alignment problems, though AI-powered warfare does get a mention. It is about the raw, structural power that flows to the builders and owners of these foundational models. The Pope argues that technology built and governed by a tiny sliver of humanity, steeped in the culture of Silicon Valley, cannot possibly serve the common good by default. Its logic is the logic of its creators, its biases are their biases, and its economic benefits flow primarily to them.
The Anthropic Connection and the ‘Babel Syndrome’
In a fascinating and very modern twist, the Pope presented the encyclical alongside Chris Olah, the co-founder of Anthropic. The inclusion of a key figure from a major AI lab, particularly one founded by former OpenAI researchers with a stated mission of AI safety, is a deliberate signal. This is not a Luddite’s rejection of technology. It is a call for a different, more thoughtful path forward, an attempt to engage the builders directly on moral and ethical grounds.
Anthropic, with its public benefit corporation structure and its focus on constitutional AI, represents an alternative ethos to the relentless commercial scaling seen elsewhere. Olah’s presence suggests the Vatican is drawing a distinction between different approaches to AI development, endorsing the one that at least speaks the language of caution and human values.
Throughout the document, the Pope returns to a powerful metaphor: the “Babel syndrome.” He describes it as an idolatry of technical prowess, a belief that we can construct a new reality through sheer computational force, leading to a breakdown in shared understanding and societal cohesion. It’s a strikingly apt analogy for an age saturated with synthetic media, algorithmic filter bubbles, and a growing inability to agree on basic facts. The fear is that AI, rather than uniting us, will give every tribe its own language, its own truth, and its own tower, leading not to heaven but to confusion and conflict.
The New World’s Land Grab
As the philosophical debate echoed from Rome, OpenAI was executing a much more terrestrial strategy. Its partnership with Grupo Folha and Grupo UOL is the latest in a string of high-profile media deals that form the bedrock of its plan to become the world’s primary interface to information.
On the surface, the arrangement seems straightforward. ChatGPT’s enormous user base, which OpenAI claims is over 900 million weekly active users, will now get summaries and content from two of Brazil’s most respected news sources, Folha de S.Paulo and UOL. The AI company frames this as a win for everyone: users get trusted information, OpenAI’s models get better, and the publishers receive attribution, links back to their sites, and, crucially, an undisclosed licensing fee.
Sérgio Dávila, Editor-in-Chief of Folha de S.Paulo, celebrated the deal, stating, “The interest of an artificial intelligence giant like OpenAI in displaying content produced by Folha de S.Paulo and UOL only reinforces the importance of professional, high-cost journalism.”
This has become the standard justification from publishers signing on. In an industry decimated by the internet’s disruption of traditional business models, a licensing deal with the new king of the internet feels like both a financial lifeline and a validation of relevance. But it also represents a profound strategic gamble.
Building the Global Content Engine
The Brazil deal is not an outlier. It is a single move in a global campaign. OpenAI has already inked similar agreements with the Associated Press, Axel Springer in Germany (owner of Politico and Business Insider), Prisa Media in Spain, and Le Monde in France. Each deal vacuums up another country’s most valuable journalistic content, feeding it into the OpenAI ecosystem to train its future models and ground the answers of its current ones.
From a technical standpoint, this strategy is brilliant. It addresses two of OpenAI’s biggest weaknesses: the tendency for models to “hallucinate” or invent facts, and the looming threat of copyright infringement lawsuits from publishers whose content was scraped from the web without permission. By licensing content directly, OpenAI gets a firehose of high-quality, factual, and legally defensible data. This data is not just used for real-time answers; it becomes part of the model’s core training, a permanent asset that improves its capabilities across the board.
But this is where the Pope’s warning about concentrated, opaque power becomes so salient. OpenAI is systematically positioning itself as the planet’s information broker. It, and a few other major AI labs like Google, are creating a new layer between the public and the primary sources of information. The terms of these deals are secret. The algorithms that select, summarize, and present the licensed content are proprietary black boxes. The decision of which publishers get a seat at the table and which are left out in the cold is made in private.
This is precisely the kind of “new dependency” the encyclical describes. Publishers, in exchange for short-term revenue, are ceding control over the distribution and presentation of their work to a single, massively powerful platform. They are becoming content suppliers for a system they do not control and cannot fully understand. This dynamic risks creating a two-tiered media landscape: those who have a deal with the dominant AI platforms, and those who do not.
The Inevitable Collision
When you place the Vatican’s encyclical next to OpenAI’s media strategy, the conflict is unavoidable. Pope Leo XIV calls for a new legal and ethical framework to govern AI for the common good. OpenAI, in contrast, is building that framework itself, not through public deliberation but through a series of private commercial agreements. One warns of opaque power, the other builds its empire on proprietary code and non-disclosure agreements.
The Pope’s critique is not that technology is inherently bad, but that its current trajectory is creating a system where power is dangerously centralized. The OpenAI media deals are a perfect case study. A single American company is now a key gatekeeper for how hundreds of millions of people in Brazil, Germany, France, and beyond will access and understand their own national news. This is a staggering consolidation of influence.
The question we are left with is fundamental. Who gets to write the rules for our shared reality in the age of AI? Will it be a slow, deliberative process involving governments, ethicists, and civil society, as envisioned in documents like Magnifica Humanitas? Or will the rules be written on the fly by the companies moving fastest, locking in their advantage through strategic deals that create a reality on the ground that regulators can only react to?
Right now, the second path is winning. While the world debates the magnificent potential of humanity, the architects of AI are busy building the magnificent machinery that will mediate it. The two conversations are happening in different rooms, in different languages. The risk, as the Pope warned, is that we end up in a new Babel, where we have more information than ever but understand each other less and less.
