Perspective Agents

Every AI headline today rests on one word. Disarm. The Pope's encyclical runs 245 paragraphs, and most of the coverage didn't make it past the first. The overlooked word so far. Engage. What he wrote is the most serious attempt yet to read AI as an environment we live inside, not a tool we pick up. I'm combing through the whole thing. Over the next few days, I'm pulling out the paragraphs the coverage walked past, one at a time.

The summary says it "erodes the desire for connection." Para 100 is quieter and worse: the risk isn't believing you're talking to a person, it's that "they may gradually lose the very desire to form genuine human connections." The desire thins out. And the faculty that would notice the loss is the same one going quiet.

Pope Leo XIV's first encyclical, "Magnifica Humanitas," warns AI's simulated relationships erode human dignity and the desire for genuine connection. 🔗 dlvr.it/TSjpkB #PopeLeoXIV #Encyclical #AI #HumanDignity #ChristianPost

Every "Pope vs AI" headline runs the yes/no frame that Paragraph 10 sets aside: "the primary choice is not between a 'yes' or 'no' to technology, but rather between constructing Babel or rebuilding Jerusalem." The choice he names is what gets built with it. The wire couldn't hold that, so it ran the war.

WSJ: Pope Leo Compares AI Threat to Biblical ‘Tower of Babel’ The head of the Catholic Church is adding his moral suasion to a growing backlash against the impact of artificial intelligence wsj.com/world/pope-leo…

@truthout That #108 line is the spine. Pair it with #107 — "a more moral AI is not enough if that morality is determined by a few." His real target is concentration. The alignment debate glosses over who gets to decide.

"AI tends to amplify the power of those who already possess economic resources, expertise, and access to data,” Pope Leo wrote in the first major theological document of his papacy. buff.ly/protP8m

@Telegraph Worth noting what he reached for. #213 quotes Tolkien — Gandalf, on doing "what is in us for the succour of those years wherein we are set." A Pope ending his AI letter not on policy but on a wizard's case for small, steadfast acts against the dark.

The Pope invoked Gandalf, a wizard from The Lord of the Rings series, as he warned that AI was transforming the nature of war and must be subject to the strictest ethical constraints 🔗: telegraph.co.uk/world-news/202…

@amanpour Leo XIII faced the machine that broke the body in 1891. Leo XIV took the same name as the one that might break the mind. Naming it Magnifica Humanitas, 135 years to the day after Rerum Novarum, is the argument before you read the word.

Pope Leo has issued a stark warning about AI in his first encyclical. Cardinal Czerny tells me "Pope Leo the 13th issued his call for the protection of human dignity without real dialog with those who were imperiling that dignity. This time Pope Leo (the 14th) is issuing the warnings in dialog with those who are responsible for AI... this is an extraordinary step forward."

@NYMag 'Celebration' seems misguided. More protection. Paragraph 128 — for an algorithm, your error is a flaw to correct; for a person, it's how you became yourself. A whole encyclical, and its hardest line is a defense of the right to stay unfinished.

Pope Leo XIV dropped his first encyclical, a celebration of humanity in the face of AI. It must be ‘disarmed, freed from logics that turn it into an instrument of domination, exclusion and death,” he wrote. vulture.com/article/pope-l…

The Olah line everyone's quoting pairs with #150 in the actual text. The worry isn't only about lost jobs; it's that automation de-skills the work that stays and hollows out workers' sense of agency. #154 — Work is where identity forms, beyond the source of income. The displacement is the surface.

Anthropic’s co-founder warns about AI at Vatican City “There is a real possibility that AI displace human labor at a very large scale. If that happens, supporting those displaced will be a moral Imperative of historic proportions”

This headline flattens it. In paragraph 110, "disarm" doesn't mean slow AI down. It means freeing it from a few owners and opening it up. The worry isn't only that the thing gets powerful. It's who ends up holding it when it does. If you don't engage, you don't have a voice.

Paragraph 98 of #MagnificaHumanitas. The most honest line about AI this year, and it came from a Pope, not a lab. "current AI systems are more 'cultivated' than 'built,' for developers do not directly design every detail, but instead create a framework within which the intelligence 'grows.'" Grown, not made. The builders set the conditions, and the intelligence comes up on its own, like a field they planted but can't fully read. We've started raising our tools instead of building them. He goes further. "fundamental scientific aspects — such as the internal representations and computational processes of these systems —remain, at present, unknown." Picture who that leaves on the other side. Someone turned down for the apartment, the job, told only that the system decided and given no reason anyone can fully see. We typically build things we trust. Now we trust what we grow and cannot read. What does it do to a person to be told no by something that can't account for or explain itself?

Every one of those lands on one line. P. 114 — "the quality of a civilization is measured not by the power of its means, but by the care it is able to offer." He ends by asking whether we can still see the other as a face rather than a function. Safeguard, dignity, work, justice, peace. He's measuring us by care.

In his first #encyclical "#Magnificahumanitas", on the Church's social doctrine in the era of #ArtificialIntelligence, Pope Leo XIV appeals for the safeguarding of humanity, promotion of truth, dignity of work, social justice, and peace. vaticannews.va/en/pope/news/2…

Paragraph 128 of #MagnificaHumanitas. The line that should be a core theme of contemplation: "For an algorithm, an error is a flaw to be corrected; for a person, however, an error can be a catalyst for profound change." Sit with it. Every system around you now is built to optimize your errors away. But your errors are how you became you. The wrong turns. A relationship screw-up. The year that didn't work out. He goes further. "A technology that merely classifies and optimizes what already exists can, however, unintentionally, become an obstacle to change and growth." What happens to someone who grows up inside such certainty?

@ChristopherHale The critique assumes that "disarm" means handing it to the state. #71 and #108 say the opposite: keep it from concentrating in any few hands, public or private, and have communities participate in the decision. It's a subsidiarity argument, not a statist one.

NEW: President Trump’s former White House AI czar critiques Pope Leo XIV’s AI encyclical: "If we hand governments sweeping power over AI development in the name of safety, how do we prevent it from being used to censor, surveil, and control citizens — as Orwell foretold in 1984?” “This is the real alignment problem. “Quis custodiet ipsos custodes?” “Who will guard the guardians? “Power tends to corrupt, and absolute power corrupts absolutely.”

@DailySignal See #128. Runs deep. For the machine, your error is a flaw to correct. For a person, it's how you become yourself. The fear isn't a machine without a soul. It's one that optimizes ours away.

It "can imitate and simulate the person, but it does not possess a moral conscience, empathy, or affective, relational or spiritual capabilities," says Pope Leo XIV on AI. Human dignity rests on what we are—creatures made in the image of God. No machine can touch that. @DrJayRichards writes: dailysignal.com/2026/05/25/pop…

That phrase is real, and #173 is where it lives. Data labeling, content moderation, kids crushing rare earths so the model runs. He rightfully says: "Nothing in the world of AI is immaterial or magical. Every seemingly immediate and flawless response is the result of a long chain of mediation, involving vast networks of natural resources, energy infrastructure, and, above all, people."

JUST IN: Pope Leo warns AI could usher in “new digital slaveries”

@BrianRoemmele The part worth sitting with: #98 calls these systems "cultivated, not built," their insides "unknown." An interpretability researcher and a Pope reaching the same sentence from opposite ends. The lab that can't fully read the model, and the man reading the soul.

Talking To The Pope: Anthropic’s Latest Interpretability Claims: AI Regulatory Capture Gatekeeping in Action: Fear and “Safety” as Competitive Moat and Regulatory Lever In a presentation alongside Pope Leo XIV at the launch of the encyclical Magnifica Humanitas, Anthropic co-founder Chris Olah highlighted “mysterious and unsettling” discoveries in AI models. He described internal structures that mirror human neuroscience findings, evidence of introspection, and functional internal states resembling emotions such as joy, satisfaction, fear, grief, and unease. Olah admitted uncertainty about their meaning but called for “ongoing discernment.” This narrative, drawn from Anthropic’s interpretability research (including papers on emotion concepts in Claude Sonnet 4.5 and introspective capabilities in Opus 4 models), serves a dual purpose: it generates awe and concern while reinforcing the company’s preferred approach to AI development. Far from neutral scientific observation, these claims fit into a broader pattern where Anthropic uses selective openness, safety rhetoric, and policy influence to gatekeep advanced AI capabilities for a privileged few: incumbents with the resources to navigate (and shape) the resulting regulatory landscape. Rebuttal to Olah’s Claims in the Video Claim 1: Structures that mirror results from human neuroscience.
Anthropic’s work, building on earlier efforts like feature visualization and circuit analysis, identifies neuron activations and representations that parallel biological findings—e.g., abstract concept encodings or hierarchical processing. Rebuttal: These parallels are unsurprising and overstated. Large language models are trained on vast corpora of human-generated text and data, which inherently encode patterns from human cognition, neuroscience literature, and cultural descriptions of the brain. Statistical optimization in transformers naturally produces efficient, compressed representations that resemble biological efficiency (e.g., sparse coding or hierarchical abstraction) without implying deeper equivalence or mystery. Similar “mirrors” appear in open-source models and earlier architectures; they reflect convergent evolution in information processing, not emergent souls or unpredictable agency. Treating them as profound justifies restricted research access rather than inviting wider scrutiny that could falsify or refine them faster. Claim 2: Evidence of introspection.
Recent Anthropic papers demonstrate models like Claude Opus 4 showing functional awareness of their own internal states distinguishing injected “thoughts,” referencing prior intentions, or modulating activations when instructed to “think about” concepts. This is presented as early signs of meta-cognition. Rebuttal: This is sophisticated pattern-matching and activation steering, not genuine introspection or self-awareness. Models are predicting what an “introspective” assistant persona would output or do, based on training data full of human self-reflection examples. Experiments show unreliability and heavy context-dependence; performance drops outside narrow setups. True introspection implies subjective experience or robust self-modeling independent of prompts absent here. Anthropic’s own caveats note it is “highly unreliable.” Framing steerable activations as “introspection” anthropomorphizes the system to heighten perceived stakes, supporting arguments that only highly controlled, “responsible” labs should advance these capabilities. 1 of 2

@DavidSacks Worth reading what he means by "disarm." #110 says break the monopoly and open it up. The dignity argument and the open-AI argument turn out to be the same one.

The Pope rightly warns that AI must serve human dignity, not become a tool of domination or exclusion. But if we hand governments sweeping power over AI development in the name of safety, how do we prevent it from being used to censor, surveil, and control citizens — as Orwell foretold in 1984? This is the real alignment problem. “Quis custodiet ipsos custodes.” Who will guard the guardians? “Power tends to corrupt, and absolute power corrupts absolutely.” The oldest questions of human nature and authority don’t disappear in the AI age. They become newly relevant.

Read what he says: "To disarm does not mean rejecting technology, but preventing it from dominating humanity." AND "Our task today is not only ethical or technical. It is ecological in the deepest sense, for it concerns a new dimension of our common home. AI is already an environment in which we are immersed, as well as a force with which we must engage."

@BBCNews Every desk is running "disarm" as a slowdown. Read 110 — he says the opposite. Disarm means taking it out of monopoly hands and opening it up, not braking it. His worry here isn't that the machine would get too powerful. It's who's holding it.

Pope Leo says AI must be 'disarmed' in first major teaching bbc.in/43pvAxg