Babel or Jerusalem: The Pope's framework for developers.

#magnificahumanitas #popeleo #ai #softwaredevelopment

During Pope Leo's presentation of his first Encyclical text last May at the Vatican, there was a word he used that made many people wonder.

Standing in front of Cardinals, diplomats, and AI top leaders (including Chris Olah, one of the founders of Anthropic), the Pope said that artificial intelligence needs to be "disarmed". He didn't mention it should be regulated, governed or supervised. He said "disarmed", and when talking about such topic and in this specific act, for sure this word was carefully chosen.

He chose the word deliberately, explaining that this moment "needs words capable of attracting attention, awakening consciences and indicating paths forward", and right after he said something else that deserves to sit with every developer and pay special attention to: the Church does not claim technical answers, but brings "a wisdom concerning the human that our present time desperately needs".

That sentence, from a man such as the Pope (a Chief of State and the leader of over 1,4 billion Catholics), addressing an audience that builds the systems in question, is not something we should take lightly. It is an invitation. And the full 245 paragraphs of his Magnifica Humanitas encyclical, on safeguarding the human person in the time of AI, are worth taking seriously not despite the fact that they come from a Pope, but in part because of it.

To this day, it's obvious that the Catholic Church has been thinking about what technology does to human beings for a very long time. It was doing it before the chips came into play, before the algorithms appeared and before the first line of code was ever written, and in fact the very same date Leo XIV chose to sign the encyclical was not accidental. It was May 15, 2026, exactly on the 135th anniversary of Rerum Novarum, the encyclical written by his predecessor and namesake Leo XIII in 1891 in response to the Industrial Revolution.

That previous Encyclical, Rerum Novarum, was surely one of the most consequential documents in the history of social thought. It was written at the exact moment when factory owners were treating workers as interchangeable units of production, when underaged children worked over 14 hours a day, when families were being uprooted from agricultural communities and deposited into urban slums with no safety net, no recourse, and no social or political voice.

Back then, the market, left entirely to its own logic, had begun optimizing for output at the expense of the people doing the work, and then Pope Leo XIII said: "...this is not acceptable, not because markets are evil, but because human beings are not inputs. They have inherent dignity. Their labor is an extension of that dignity, not a commodity. The state, employers, and institutions all carry obligations that the profit motive alone cannot be trusted to honor".

It took decades for those ideas to fully materialize in labor law, workers rights, and social policy, but they eventually did and the world is now measurably better for it. The wisdom basically arrived before the legislation, and the legislation needed the wisdom to know what it was trying to protect.

As it happened to the Pope back then, now Leo XIV looked at AI in 2026 and saw the same shape of problem, moving forward to elaborate a very well built and extremely symbolic Encyclical where the organizing metaphor at its heart is striking in its clarity.

The encyclical frames the choice before humanity not as "AI yes or no" but as Babel or Jerusalem, with Babel meaning centralized power, optimizing for its own expansion, indifferent to the human cost and building something that ultimately fragments rather than connects. Jerusalem, in the other hand, means a shared reconstruction centered on dignity and oriented toward the common good, building something that makes human community possible rather than undermining it.

In any case, the Encyclical does not say AI is Babel, but what it says is that AI can turn into either, and that which one it becomes depends entirely on the choices made by the people building it, funding it, deploying it and governing it right now. Because according to the Pope, the technology is not neutral and it carries the values (or the absence of values), of those who made it. This is the concern that runs through every chapter.

The Encyclical names aspects of AI like job insecurity, manipulation of information, privacy violations, ideological bias and autonomous weapons as specific risks, but maybe the deeper risk underneath all of them is that the acceleration of AI deployment could be outpacing the moral and social infrastructure needed to shape it toward human ends. That the decisions being made this year, mostly by a small number of people and at extraordinary speed, will have consequences that fall most heavily on people who had no voice in making them.

Leo XIV warned specifically that rapid automation could displace workers and reshape labor markets in ways that risk leaving many in "forced inactivity" undermining both human dignity and social stability, in a clear echo of Rerum Novarum that is hard to dismiss as just a coincidence.

So once this is explained, and looking deeper into what this actually means for those of us building with AI, it must be noticed that most developers working on AI products are not trying to build Babel in purpose but instead they are just trying to solve real problems by using the outstanding advantage AI brings. They care about what they make, but the question Magnifica Humanitas puts to them is not if they are good or bad guys, but something much harder: Are the systems you're building designed to answer to the people they affect?

That is the question and this question matters in practice, not just in theory, because an AI system that optimizes for engagement metrics without accounting for the psychological wellbeing of users is not necessarily malicious, but it's just indifferent to a variable it was never asked to consider.

An AI hiring tool that produces discriminatory outcomes isn't evil, it's just a model that learned patterns from data that encoded historical inequity, and nobody built in a correction. An AI system deployed in a low income country that was trained entirely on data from wealthy ones is not necessarily cruel, but it's just misaligned in ways that weren't caught before deployment, and the people who pay the price for that misalignment had no seat at the table when it was designed. And so on.

These are engineering and governance problems and according to what the Pope says they also have engineering and governance solutions. The Pope is not asking developers to pray over their pull requests. He's asking the industry as a whole to take seriously a question that market incentives alone cannot answer. He asks the industry to think about who are their AI's designed for, who might it harm and what would we need to build differently if those people's wellbeing actually constrained our design choices...

The Encyclical's framing is clear and it points out that the challenge is not simply to manage technology, but to shape it with conscience so that it truly serves life, not just the ambitions of a few, and that's not a theological position but it should be a design principle. And it's one that many developers who will be most proud of their work in twenty years time from now are already trying to live by.

Going back to the Industrial Revolution's Previous Encyclical, written in 1891, the wisdom in Rerum Novarum was considered idealistic by the people with the most power to ignore it. Brutal things like child labor laws came anyway, the eight hour workday came anyway, social safety nets came anyway, the market didn't voluntarily generate those protections and the moral argument had to be made clearly enough and persistently enough that it eventually became politically irresistible.

In our own current case, we are still early in that process with AI, the legislation is fragmented and behind the technology, the governance frameworks are still forming, the power is highly concentrated and the people most affected have the least influence. So in line with the previous experience with Rerum Novarum, what Magnifica Humanitas offers is basically a framework for thinking about what we're actually trying to protect before the specific rules get written. It's a reminder that human dignity is not a variable to be weighted against efficiency, that the common good is a real constraint (not a marketing strategy), and that technology shaped without conscience produces outcomes that conscience, arriving later, will spend decades trying to repair.

Developers are not just spectators to that process. They are, in fact, right now making the decisions that will determine which side of the Babel / Jerusalem choice we end up on.

The Pope showed up in person to say so. The co founder of one of the largest AI companies was in the room to hear it. Both of those things, happening simultaneously, feel like a moment worth paying attention to, and we should all create our true opinion.

Sources: