AI Is Not the Problem. Who Controls It, Is.
The real risk of artificial intelligence is not technological but political. Between the papal encyclical Magnifica Humanitas and Silicon Valley's responses, an ancient dilemma resurfaces: who controls the controllers?
When Leo XIV signed the first encyclical dedicated to artificial intelligence, Magnifica Humanitas, many expected a technical document, perhaps a list of ethical prohibitions, something resembling a bureaucratic circular with an apostolic blessing. Instead, the Pope did something more radical and more uncomfortable: he talked about power. He wrote that artificial intelligence must not become a new Tower of Babel, a system built to concentrate economic, cognitive and political dominion in the hands of a few. He did not condemn it. He warned it. And the difference, in this case, is everything.
The response came quickly. David Sacks, former White House AI and cryptocurrency adviser during the Trump administration and a significant figure in the Silicon Valley ecosystem, replied with a question that sounds like an elegant counterattack: agreed, protecting human dignity is right, but if we hand governments excessive regulatory power over AI in the name of security, who will protect us from censorship, surveillance and citizen control? Who will prevent those very tools of protection from becoming instruments of oppression?
This is where the debate stops being an academic confrontation and becomes something older and more urgent. Sacks did not cite hypothetical cases: he evoked Orwell, 1984, the principle that every technology born to protect can transform into a mechanism of control. And history gives him enough material to make that case. Mass surveillance, social credit systems, profiling algorithms used by authoritarian governments around the world, from China to Russia to less visible regimes, are all built on infrastructures that, at least on paper, were born to guarantee order and security.
What strikes you, however, is an almost paradoxical coincidence: the Pope and Sacks are talking about the same problem. Both fear the concentration of power. Only one points the finger at big tech, the major platforms, the multinationals accumulating data and computational capacity without answering to any democracy. The other points the finger at states, those governments that could use regulation as an alibi to control information and silence dissent. Two identical diagnoses, two different patients.
According to the Artificial Intelligence Index 2025 report published by the Stanford Human-Centered Artificial Intelligence Institute on 7 April 2025, the number of national AI regulations has more than tripled in the last four years. This means that AI governance is already underway, for better and for worse, often without global consensus and often with logics that reflect the interests of those who legislate, not those who are governed.
And this is where the most unsettling point of all emerges, the one that neither the pontiff nor the technocrat has resolved: the problem of AI alignment is, at its core, a human problem. For years, experts have debated how to align artificial systems with human values, how to ensure a machine acts ethically, predictably, safely. But the question that remains suspended, embarrassing and unanswered, is a different one: have we ever managed to align human beings with each other? Have we ever reached a stable agreement on what just power means, legitimate control, security without oppression?
Artificial intelligence has made that question impossible to ignore. Not because machines are dangerous. But because they are handing human beings, still unresolved, an amount of power we have never had before. And whoever does not ask themselves that question is, in all likelihood, already part of the problem.
Sources: Papal Encyclical Magnifica Humanitas, Holy See, 2026; Artificial Intelligence Index Report 2025, Stanford HAI, 7 April 2025; public statements by David Sacks, former White House AI & Crypto Advisor.