A 40,000-Word Document With a Tech Industry Audience in Mind
Pope Leo XIV published his first encyclical on Tuesday in Rome, a 40,000-word document called Magnifica Humanitas – Latin for “Magnificent Humanity” – and the person standing next to him when he did was not a cardinal or a head of state. It was the co-founder of Anthropic, one of the most prominent AI safety companies operating today. That detail alone signals who this document is talking to.
The encyclical’s central demand is that artificial intelligence be “disarmed.” Leo acknowledges the word is strong. He chose it deliberately, writing that “this moment needs words capable of attracting attention, awakening consciences, and indicating paths forward for humanity.” AI, in his framing, must be “freed from logics that turn it into an instrument of domination, exclusion, and death.”
Where the Document Lands Its Hardest Hits
The encyclical does not spend its 40,000 words on vague philosophical anxiety. It names specific categories of concern and does not soften them. AI-powered autonomous weapons receive direct, uncompromising criticism – the kind of critique that arms manufacturers and defense contractors funding AI research will find difficult to dismiss as merely theological. When a document of this institutional weight calls out a product category by name, the industry notices.
Leo also targets what he calls neo-colonial attitudes in data collection – the practice of extracting behavioral and personal data from populations in lower-income countries, feeding it into models that generate profit elsewhere, and returning nothing of value to the source. This is not a new critique in tech-policy circles, but embedding it inside a papal encyclical gives it a very different kind of distribution network than an academic paper or a Brussels regulatory filing.
Then there is the property question. The document specifically calls out the hoarding of “new forms of property, such as patents, algorithms, digital platforms, technological infrastructure, and data.” That list reads less like theology and more like a index of Silicon Valley’s most guarded assets. Patents on foundational AI methods. Control over the platforms that billions of people use daily. Ownership of the infrastructure – the data centers, the undersea cables, the GPU clusters – that the entire global internet now depends on. Leo is not asking those who hold these assets to share nicely. He is framing concentrated ownership as a moral failure.
The Gandalf Reference and What It Actually Means
Buried inside a document this long is a citation from J.R.R. Tolkien’s The Lord of the Rings – specifically, Gandalf. The pope invoked the wizard in making his case about power and restraint. The reference is not decorative. Gandalf’s defining characteristic in Tolkien’s story is his refusal to wield the One Ring even when he could – because he understands that absolute power, even in the hands of someone with good intentions, warps toward domination. Applying that logic to AI development is not a subtle move.
The choice to quote Tolkien rather than, say, a UN resolution or a peer-reviewed ethics framework is worth noting. It speaks to an audience that is not necessarily religious – engineers, product managers, founders – in a language that cultural memory has already made legible. Whether that framing lands as wise or as gimmicky depends entirely on who is reading it.
Anthropic’s Presence and What It Does Not Confirm
The attendance of Anthropic’s co-founder at the encyclical’s release is conspicuous. Anthropic was founded specifically around the premise that AI development carries existential risk and requires careful safety-oriented research – a founding philosophy that shares at least surface-level common ground with the encyclical’s concerns. The company has positioned itself as the responsible actor in a field it acknowledges is dangerous.
But standing next to a pope at the launch of a document demanding AI “disarmament” does not mean Anthropic endorses every clause. The company builds large language models and charges for access to them. Its products are commercially deployed, its investors expect returns, and its researchers are in direct competition with the labs the encyclical implicitly criticizes. The optics of the moment are notable. The operational implications are less clear.
What the presence does confirm is that the conversation around AI governance has now moved well beyond regulatory agencies and parliamentary committees. A 40,000-word moral document from the Catholic Church – an institution with over a billion members globally – entering the debate on autonomous weapons, data colonialism, and intellectual property concentration adds a layer of pressure that is structurally different from a government fine or a congressional hearing. It is not enforceable. It is, in a certain sense, harder to ignore than enforcement.
The encyclical’s language around “disarmament” is aimed squarely at making that difficulty felt. Whether the industry responds with engagement, dismissal, or the polished non-answer that tech has refined into an art form, the document is now in circulation – 40,000 words long, written with a Tolkien reference in the middle, and introduced to the world with one of AI’s biggest players standing in the room.
The next Anthropic product announcement will carry that image behind it, whether the company wants it to or not.
