The biggest AI companies in the world were asked what they thought about the Pope’s first encyclical on artificial intelligence. Amazon, Google, Meta, Microsoft, OpenAI, Nvidia, xAI: almost all of them declined to comment or didn’t respond to press requests at all.

Sam Altman said nothing. Elon Musk said nothing. Mark Zuckerberg said nothing.

That silence is the most interesting thing about the whole story. It’s evidence of something Magnifica Humanitas names directly: there’s no shared vocabulary yet for questions of AI moral architecture. And it’s what I kept returning to in my conversation on RTHK Radio 3 this Friday morning.

What the Document Is

Magnifica Humanitas is Pope Leo XIV’s first encyclical, signed on 15 May 2026. It’s 42,300 words, it’s not anti-AI, and it’s not a prohibition on anything. It’s a moral standard: AI systems must serve human dignity rather than concentrate power, and the person at the end of any system has dignity the system is required to respect. For the full historical analysis, including the deliberate 135-year parallel with Rerum Novarum, I wrote about it earlier this week.

The Press Split: Which Frame Are You Using?

When a document like this lands, the way different outlets cover it tells you which frame they’re applying and which audience they’re writing for.

The Guardian framed it as a “humanity first” message the secular world can get behind: accessible, non-threatening, a bridge between religious and secular concern. The Wall Street Journal called it “a text poised to define Leo’s papacy” and a long-awaited moral anchor for policymakers. The BBC focused on the Pope’s direct challenge to those in power. Forbes took the Tower of Babel imagery and wrote about “moral, social collapse.”

The sharpest observation came from TechCrunch, which noted that “AI is the hook, the problems Leo focuses on are older and more pervasive.” That’s correct. This is a document about who controls the systems that shape people’s lives. AI is the current expression of a much older question about political economy and the structure of authority.

Each of those framings is accurate in its own way. But together they reveal something: there’s no shared vocabulary yet for what the encyclical is actually doing. Every publication reached for whatever frame it already had. The result is multiple partial readings rather than a single coherent understanding of what was named.

That fragmentation is itself evidence of the gap the document identified.

Why the Silence From Silicon Valley Tells You Everything

Here’s the question that matters: are the people actually building AI paying attention?

The answer, in most cases, is silence.

Sam Altman, Elon Musk, Mark Zuckerberg: nothing publicly. Amazon, Google, Meta, Microsoft, OpenAI, Nvidia, xAI: declined to comment or didn’t respond to press requests. The only notable exception: Christopher Olah, co-founder of Anthropic, was in the room at the Vatican when the document was presented.

I don’t think this silence is indifference. I think it’s something more revealing: the AI companies have no vocabulary for this kind of challenge.

When a technical claim is made about AI, “this model is dangerous,” “this benchmark is misleading,” “this capability is overstated,” the AI companies know how to respond. They dispute the evidence, produce counter-evidence, publish research, hire teams. They’ve built extensive machinery for technical disagreement.

When a regulatory claim is made, “this needs governance,” “this should be licensed,” “this poses systemic risk,” they also know how to respond. They meet with legislators, produce policy papers, propose alternative frameworks, build government affairs teams. They’ve built machinery for political engagement.

But Magnifica Humanitas is neither a technical claim nor a regulatory proposal. It’s a moral naming. It’s an institution with 2,000 years of experience thinking about what human beings owe each other, arriving at a question: what is this technology permitted to do to the human experience? That isn’t a peer-reviewed paper to be disputed on technical grounds. It isn’t a piece of legislation to be lobbied against. The major AI companies have no ready-made response to moral authority of this kind.

They’re silent because they don’t know what to do with it.

That silence isn’t incidental. It’s evidence of the gap the document identified. When moral architecture arrives at a question and the people most implicated by it have no response, the naming has done its work. The institutions will catch up. They always do.

That Anthropic sent a co-founder to the Vatican is, I think, the single most significant corporate signal in this entire story. It says: we know we need this conversation. Whether the rest of the industry follows is worth tracking carefully.

The Critics Who Got It Wrong

Two critics went public with substantive objections.

A venture capitalist named Eddy Lazzarin called the document “light on the theology of artificial intelligence.” AI researcher Pedro Domingos said it was based on “ignorant and wrong-headed things about AI.”

Both critiques attacked the wrong thing. And in doing so, they revealed exactly what the encyclical was pointing at.

Magnifica Humanitas isn’t making technical claims about how AI works. It isn’t asserting that large language models are conscious, or that transformers have moral weight, or that current systems meet any threshold of agency. The document is making moral claims: what AI is permitted to do to people, to their dignity, their agency, their capacity to make meaningful choices about their own lives.

Critiquing this document on technical accuracy is the equivalent of critiquing Rerum Novarum for not understanding metallurgy. The 1891 encyclical didn’t need to understand how coal extraction worked in order to say that children shouldn’t be made to work in coal mines. The mechanism wasn’t the point. The person was the point.

What Lazzarin and Domingos revealed, in applying a technical frame to a moral document, is that they also lack the vocabulary. They went to the only frame available to them, technical accuracy, because that’s how the technology sector has trained itself to evaluate claims. Focused almost entirely on what can be done. Almost entirely silent on what should be permitted to happen to the people inside the systems being built.

This isn’t a personal failure. It’s a structural condition. And the encyclical is pointing directly at that structure.

The EU’s Response: Mistaking Regulation for AI Moral Architecture

The European Union has moved faster on AI regulation than any other governing body in the world. And their initial response to the encyclical was essentially: we already have the AI Act, we’re fine.

That response is, in some ways, the most instructive of all.

The EU has done something genuinely significant with the AI Act: built legal architecture that categorises AI systems by harm potential, requires transparency and human oversight in high-stakes contexts, and creates accountability structures that didn’t exist before. That’s real work. It matters.

But the encyclical is asking a different question. Legal architecture tells you what’s permitted. Moral architecture asks whether what’s permitted is right. Those aren’t the same question. Confusing them is a form of blindness that boards, executive teams, and governments of all kinds need to examine honestly.

The fastest-moving regulatory body in the world responded to a moral challenge by citing its own legislation. That’s the defensive move of an institution that has confused legal compliance with moral integrity. If you believe building the right rules is the same as answering the right question, you’ll always be behind the people asking the deeper question.

This matters well beyond Brussels. Any board, any executive team, any leader who believes complying with AI regulation means they’ve addressed the moral question is making the same error. They’ve addressed the legal question. The moral question is still waiting.

What This Means for Professionals Living Inside AI Systems

Every professional listening to RTHK this Friday morning has made decisions, or had decisions made for them, by AI systems in the past year. Hiring decisions. Credit assessments. Health information. News feeds. Insurance pricing. Access to services. In almost every case, they didn’t ask for those systems. In almost every case, they don’t know how those systems work. And in almost no case were they told: this system is making a judgment about you, and here’s the moral framework that governed its design.

The encyclical is naming something those people have been feeling without language for. Not the technology itself. The absence of architecture around it.

In my research into Australian and US organisational leaders, I found that zero per cent of government and policy executives trusted AI with capital spend decisions. Not low confidence: zero. That number isn’t a failure of AI capability. It’s a Trust Cliff: the point at which the absence of moral architecture makes the risk of proceeding feel greater than any possible gain. People can feel that the infrastructure isn’t there, even when they can’t name what’s missing.

What’s missing isn’t more technical capability. It isn’t more regulation, exactly. It’s a clear, shared answer to the question: what is this system permitted to do to me as a person?

That question is now officially in the room.

What Phil Asked Me On Air

There’s a reason Phil’s question landed the way it did. He asked what most people would ask: does any of this actually change anything?

My answer: not immediately. And that’s the point.

I covered the full 1891 lesson in detail this week. The short version: Rerum Novarum didn’t close a single coal mine the day it was published. But it named something that couldn’t be unnamed after that. The laws and unions and safety standards that followed over fifty years were its ripple effects. Always in that order.

The same logic applies now. Magnifica Humanitas won’t close a single data centre. It won’t stop any AI company from building what it’s building today. But it will enter the governance conversation: through every regulator who needs moral authority for their position, every board director who’s been circling this question looking for language, every professional who’s been affected by an AI-driven decision and wants to say: this isn’t only a technical matter. It’s a matter of what I’m owed as a person.

If you’re reading this looking for immediate legislative consequence, you’re reading it on the wrong timescale. Read it on the 1891 scale. The ripple effects are beginning. You’re watching the first stone land in the water.

What to Watch Now

I’ve written before about the hidden strategic risks most leaders aren’t yet tracking. The AI moral architecture gap belongs on that list, and it’s moving faster than most organisations realise.

Watch the AI companies: not their public statements, which will be carefully managed. Watch whether “moral architecture” or language about what AI is permitted to do to people begins appearing in their internal frameworks or public commitments. Watch whether any other AI leader follows Olah’s example and shows up to conversations of this kind.

Watch the regulators: specifically whether anyone begins to distinguish, in their public reasoning, between legal compliance and moral integrity. When that distinction appears explicitly in regulatory language, the next generation of AI governance is beginning.

Watch the organisations in your own sector. The boards and executive teams already asking “what should our AI systems be permitted to do to the people inside them?” are building something the rest of their industry will eventually need to match. That’s how moral architecture spreads: not by mandate, but by example.

And watch your own organisation. If you’ve deployed AI systems that make decisions affecting employees, customers, or citizens, and you haven’t asked this question explicitly, ask it now. Not because the law requires it yet. Because the question is in the room, and it will be answered. The only choice is whether you answer it first, on your own terms, or later, on someone else’s.

The threshold has been crossed. The question has changed.

Not what can AI do. What should it be permitted to do to the human experience.

Choose Forward.

If tracking signals like these is part of how you prepare, Signal Room is the practitioner intelligence service I built for exactly that. Weekly provocation cards, monthly analysis, direct access.

Get Glimpses from the Future three times a week on LinkedIn: signal analysis for leaders navigating uncertainty.

This post is based on my segment with Phil Whelan on RTHK Radio 3 Morning Brew, Friday 29 May 2026.

You can listen to the full segment below:

Morris Misel
Foresight Strategist | Keynote Speaker | Advisor
morrisfuturist.com

#MorrisMisel #MagnificaHumanitas #AIGovernance #AIEthics #MoralArchitecture #ImmediateFutures #FutureOfWork #Leadership #RippleEffects #StrategicForesight #HUMAND #TrustCliffs #KeynoteSpeaker #ForesightStrategy #ArtificialIntelligence