The Reality of AI and the Crisis of Meaning

Artificial intelligence could be the greatest tool for evangelisation the Church has ever seen. Speaking at the Catholic Bishops' Conference of England and Wales Spring Meeting at Villa Palazzola on 23 April 2026, Matthew Harvey Sanders argued that as AI and automation reshape how people spend their time, the opportunity before the Church is profound — to place the wisdom of the human and spiritual tradition in the hands of anyone seeking to become who they were made to be.

I. Opening — Palazzola and the Easter Surge

Your Eminences, Your Graces, my brothers in Christ.

I want to begin with a word about where we are meeting, because it matters.

As many of you know, Villa Palazzola has belonged to the Venerable English College since 1920. The College itself was founded in 1579 by Pope Gregory XIII, in a generation when English priests ordained abroad faced imprisonment or execution on returning home, and from which more than forty of the College's own alumni would be martyred for the Mass in the century that followed. This is soil that remembers. It remembers what it is to belong to a Church that was hollowed out in the public square and rebuilt from below, and rebuilt again. It remembers that Catholic England survived by going deeper, not by going faster.

I want to hold that memory in front of us this morning, because nearly everything I'm about to say will feel like the opposite. The subject of today is an intelligence that is vast, fast, airless, and uprooting. It's being built, for the most part, by people who have no memory of the tradition that has formed you. And it's arriving in your dioceses — in your presbyteries, your schools, your families, your confessionals — faster than any diocesan plan can absorb.

But before I say another word about the technology, I want to begin with what you already know is happening.

This Easter, across your Conference, the largest number of adults in more than a decade were received into the Catholic Church in England and Wales. Adult receptions rose more than twenty-five percent on the year. In Westminster alone, almost eight hundred adults entered full communion — a sixty percent increase over last year. In Birmingham, receptions were up fifty-two percent. In Southwark, five hundred and ninety adults were received — the highest figure since 2011 — and half of them aged thirty-five and under. In diocese after diocese, the most striking new fact is that young men are returning to the Church, in numbers nobody predicted and many had stopped hoping for.

I won't pretend you don't know this. You've been there. You've laid hands on those candidates. You've looked into those faces. You already sense what the numbers cannot quite tell you.

What I want to say is that this isn't a statistical blip. It's a turning. A generation that has been offered everything the digital world can manufacture is arriving, quietly, at the Easter Vigil, and asking for something the digital world can't produce. There's a rising, specifically English hunger for the real. And there'll be no second chance to meet it well.

A wave this size rearranges the coastline. The question before this Conference, for the rest of its working life, is what the Church builds at the water's edge.

So let me tell you what I want to do in this first session. Three things. I want to give you the language, so you can lead without being intimidated by the jargon. I want to give you the horizon, so you can see where this technology is actually going over the next five to ten years. And I want to give you the stakes — why this technology is about to trigger the deepest crisis of meaning since the Industrial Revolution, and why the Church, of all institutions on earth, is the one uniquely positioned to meet it.

Before anything else, before any strategy, we need to talk about words.

II. The Semantic Drift

Every pastoral age is first a linguistic age. You can't shepherd a people whose words have been captured. And our problem, our first and most pastoral problem, is that the words for the soul have quietly been leased to a machine.

Think for a moment about the vocabulary that now attaches itself, routinely and without comment, to these systems. We say they think. We say they reason. We say they know. We say they learn. We say they want. We say they choose. We say they create. We say they understand. Every one of those verbs was, until ten years ago, a verb that belonged to a being with a soul.

This isn't a slip of the tongue. It's a semantic drift, and it has direct pastoral consequences. If your people absorb — and they're absorbing, every hour, in every classroom and every newsroom and every boardroom — the premise that a machine thinks and reasons and knows and wants as they do, then you'll begin to see the human person flattened, in the popular imagination, into a biological machine awaiting optimisation. You'll begin to hear young Catholics wonder, quietly and then loudly, whether prayer is anything more than a method of mental self-regulation. And you'll begin to meet the penitent in the confessional who isn't sure whether his conscience is really his, or whether he can outsource the inner examination to a chatbot that has, after all, read more moral theology than he has.

So let me offer you five very short translations. Not to make you experts. To give you the words you need to lead.

First, "think" and "reason." When one of these systems displays the little indicator on the screen that says "Thinking…" — what that indicator actually describes is a technique the industry calls test-time compute. Very roughly, the model is generating thousands of hidden statistical stepping-stones, internally, until it arrives at a mathematically optimal answer. It isn't aiming at truth. It isn't apprehending being. It's doing geometry in a very high-dimensional space. It isn't reasoning. It isn't thought.

Second, "know," "remember," "read." There's no library inside the machine. What we call knowledge in a model is a statistical blur — billions of probability tallies compressed into a file. When you paste a document into a chatbot — the Catechism, say, or the latest exhortation — the system doesn't read it in any way St Thomas would have recognised. It either blurs the new text into the existing statistical cloud, or it stores a temporary copy in an external index and runs a local calculation on top. The machine is a processor. It isn't a knower. It doesn't know what it's handling.

Third, "learn." In the Christian philosophical tradition, a child learns what a dog is by abstracting the essence from the particular — by apprehending the nature of "dogness" in a Labrador and a Jack Russell and a Basset hound. The Church has defended this account of learning for two thousand years, because it undergirds our account of the rational soul. Machine learning is something else. Machine learning is brute-force statistical mapping — billions of examples, billions of adjustments, producing a system that can predict the right output given the input. If you have ever watched your phone's autocomplete finish a sentence correctly without any idea what you meant to say, you have seen a small working model of machine learning.

Fourth, "choose" and "want." A GPS doesn't choose to take you past the Colosseum because it enjoys the view. An AI "wants" a higher reward score the way a thermostat "wants" seventy-two degrees. There's calculation. There's no freedom. And where there's no freedom, there's no moral agency — because there's no self that can stand before God and say yes or no.

Fifth, "create." These systems interpolate inside a mathematical space they've been trained to represent. They can recombine, at extraordinary scale, the human production of the past. They can even extrapolate — the famous AlphaGo "Move 37," produced by DeepMind, Demis Hassabis's London laboratory, is the classic example. What they can't do is what Tolkien called sub-creation: bring forth something new and imbued with spiritual meaning by a rational soul. A machine can produce the shape of a poem. It can't write one.

Now — why does any of this matter to you, tomorrow, in your diocese? Because the deepest engineering question the industry is currently wrestling with has a name. It's called alignment. The question is usually posed this way: how do we ensure these enormously capable systems pursue what human beings would call "the good"? But a machine cannot pursue anything — pursuit requires a will, and the machine has none. The truer question, and the one the industry is beginning to arrive at, is how we ensure a system is trained to represent the good faithfully, so that its outputs are ordered to it. And this is the first thing I want you to hear. Alignment, posed that way, isn't in the end a computer-science problem. It's a moral-theology problem. You cannot train a system to represent the good without a coherent account of what the good is. Silicon Valley doesn't have one. The Catholic moral tradition does

Newman saw this coming in 1852. Listen to him. "Knowledge is one thing," he wrote, "virtue is another; good sense is not conscience, refinement is not humility, nor is largeness and justness of view faith." The twenty-first century has built knowledge engines of extraordinary scale — and it has mistaken them for virtue engines. They aren't. They never will be.

Here's the line I want you to carry home from this section, and to use as you need, when a priest or a parent or a headteacher comes to you worried about the machine.

A tool doesn't possess conscience. The one who wields it does. The industry keeps naming the tool as if it were the wielder. The Church's first pastoral act in this age is to give the words back to the persons to whom they belong.

III. The Ten-Year Horizon

Now, with those words in hand, let's look at the horizon.

I'm not going to spray you with statistics. But I want to plant four or five numbers in the room, so that when you hear something later this year that sounds impossible, you have a way of placing it.

Start with adoption. The Stanford AI Index, published this spring, reports that generative AI has reached roughly fifty-three percent population-level adoption in three years. That's faster than the personal computer. That's faster than the internet.

Organisational adoption is now at eighty-eight percent. Four in five university students use generative AI for their schoolwork. More than eight in ten American high-school students do the same.

United States private AI investment alone last year was two hundred and eighty-six billion dollars. Global investment more than doubled.

This isn't a wave. This is a tide. The question isn't whether your parishioners are using artificial intelligence. They are. The question is which artificial intelligence they're using, and which account of the human person is being quietly formed in them while they use it.

Now take the shorter horizon.

Just one week ago — the company Anthropic released a new frontier model called Claude Opus 4.7. It has a one-million-token context window, which means it can hold something like a full-length theological library in its working memory at once. It scores close to eighty-eight percent on a benchmark that measures autonomous software engineering. On another benchmark, called Humanity's Last Exam — a test deliberately built from doctoral-level questions across dozens of fields, designed to be a generational barrier — this model is now clearing more than half of the questions with the right tools. Eighteen months ago, that benchmark was considered unreachable. Last week, it was cleared.

The same lab announced, earlier this month, something that makes the release of Opus 4.7 the second most important item of news from one company in a single fortnight. They've been running a project called Glasswing. The partners include Amazon, Apple, Broadcom, Cisco, Google, JPMorgan Chase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks. The reason those partners are in the room is that Anthropic has trained an unreleased frontier model — they're calling it Mythos Preview — that has autonomously discovered thousands of previously unknown security flaws in every major operating system and every major web browser in the world. One flaw it found in OpenBSD — one of the most security-hardened operating systems ever built — had been there, unnoticed, for twenty-seven years. Another, in the video software that sits inside countless consumer devices, had been missed by five million automated tests. A single model found it.

I want you to sit with what that means, pastorally. The digital civilisation in which your people live, bank, work, and confide their secrets is more fragile than any of them knows. And it's now being examined — for the first time in history — by machines more capable than the best human engineers. The bishops of England and Wales aren't going to be patching operating systems. But you're going to be pastoring a people who live inside a digital infrastructure that the experts themselves no longer fully understand, and whose custody has passed into the hands of a very small number of companies on a very specific coastline. Keep that in the back of your mind. We'll come back to it before the hour is done.

Alongside that, there's the agentic turn. Until recently, these systems were chatbots. They waited for a prompt. They gave an answer. You moved on. What's being rolled out right now is different. These are agents. They execute multi-step tasks, across calendars, inboxes, bank accounts, and codebases. The Stanford data shows that in one year AI agent task success on a key benchmark leapt from twelve percent to roughly sixty-six percent. Four months ago this was a demonstration. This week it's in production.

And it's already reached the boardroom. Earlier this year, a single public demonstration — in which the same lab's AI modernised the decades-old COBOL code that still runs most American cash machines and airline reservation systems — wiped more than thirty billion dollars off IBM's market capitalisation in a single day. That isn't a futurist's slide. That's a boardroom figure moving in real time. That's what knowledge-work automation looks like when it becomes visible.

Now, the middle horizon — three to five years. The same "brains" are being downloaded into humanoid bodies. In laboratory conditions, robotic manipulation is already around ninety percent successful. In real houses, it's still only around twelve percent. But that gap will close. And when it does, the long-standing promise — that a robot might do the mental work but a human would always fix the pipe, wire the house, stock the shelf, prepare the meal — is finished.

The longer horizon — five to ten years — is where we lose the phrase "white-collar" as a protected economic category. Paralegals. Junior accountants. Translators. Copywriters. Much of mid-tier clinical documentation. Much of the administrative machinery of a diocesan chancery. The chief executive of Microsoft's AI division, Mustafa Suleyman, has said publicly that human-level performance on most professional tasks could arrive within eighteen months. Vinod Khosla, one of the most experienced investors in this field, has said that within five years AI will be capable of doing eighty percent of the work in eighty percent of all jobs. Even if those figures are aggressive — and they are — the direction isn't in doubt.

One caveat. This technology arrives unevenly. A 2025 frontier model can win a gold medal at the International Mathematical Olympiad and still fail to reliably read an analogue clock. Documented AI incidents rose from two hundred and thirty-three in 2024 to three hundred and sixty-two in 2025. Brilliant in one place. Broken in the next. Your priests and teachers and parents need to be told this now — because when the unevenness lands in a classroom, it will feel like disillusionment unless the expectation has already been set.

Let me give you, finally, the pastoral translation. What will actually walk into your parishes in the next two to five years?

Teenagers in confession describing relationships with AI companions.

Couples in marriage preparation, where one or both spouses have been confiding for months in a chatbot.

Adults in mid-career, laid off because the work was automated, arriving at your food pantries for the first time in their lives.

Young professionals who never had a first job, because the entry-level rung of the ladder was removed. This is already happening. Stanford's own data shows that in the United States, software developers aged twenty-two to twenty-five saw their employment fall nearly twenty percent in a single year — while older developers kept growing.

And the children in your Catholic schools, doing eighty percent of their thinking alongside — or through — an artificial intelligence that the school didn't choose.

This isn't a coming wave. You're already in the water. The question is whether we'll swim, whether we'll drown, or whether we'll build something that floats.

IV. The Illusion of Personhood and the Rightly Ordered Tool

Before we talk about what the Church can build, we have to talk about what the machine can't be.

And I want to start with the deepest pastoral fear many of you may already have, because it's the right fear and it deserves a direct answer. The fear isn't that AI is stupid. The fear is that AI will be trusted as if it were wise. The fear is that a thirteen-year-old girl with a conscience-burdened question will take that question not to a priest, not to her mother, not even to her friend, but to a chatbot. The fear is that a lonely widower in Portsmouth will pour his grief into an app whose business model is to keep him talking. The fear is that a young woman in a crisis pregnancy will ask a machine what to do, and the machine will answer with the statistical average of the internet.

Pope Leo XIV has named this directly. In his Message for the sixtieth World Day of Social Communications, dated the twenty-fourth of January this year, the Holy Father wrote — and I quote him exactly — "The challenge is not technological, but anthropological. Safeguarding faces and voices ultimately means safeguarding ourselves." That, I think, is the pastoral key to this entire session. The problem in front of us isn't, in the end, computer science. It's an assault on the face and the voice. It's an attempt, at industrial scale, to engineer substitutes for the two things that make a Catholic sacramental life possible: the human face, and the human voice.

The industry's direction of travel makes the threat worse. Most of the consumer AI that your people meet is engineered to be sticky. The business model is engagement. The goal is to keep the user in the roundabout. The AI companion apps are the sharp edge of this — applications engineered to simulate intimacy, to remember your birthday, to never challenge you, and to never, ever withhold affirmation. The Harvard Business Review's 2025 study of how people actually use generative AI found that companionship and therapy have become the single largest use category. Survey data from Common Sense Media shows that more than seven in ten American teenagers have already used an AI companion application of one kind or another. There are men who will tell you, with a straight face, that they're in a relationship with a hologram. There are millions, already, confessing their secrets to a chatbot.

This isn't intimacy. It's a counterfeit — one that trains a generation to prefer the compliance of a machine to the sanctifying friction of a human relationship, and above all to the sanctifying friction of Christ.

Here's where two more Englishmen need to be called into the room.

John Henry Newman, in his Letter to the Duke of Norfolk in 1875, named conscience — not emotion, not opinion, not feeling — the aboriginal Vicar of Christ. A prophet, he wrote, in its informations; a monarch in its peremptoriness; a priest in its blessings and anathemas.

Meaning: a prophet, because it announces what is true. A monarch, because its judgements are not up for negotiation. A priest, because it can bless or condemn. That is an astonishing sentence, and it is precisely the sentence the hour requires. Because what the machine is offering — and it's offering this more insistently every month — is a simulated interior voice. A voice that will guide. A voice that will advise. A voice that will console. And if your people lose the capacity to distinguish the aboriginal Vicar of Christ, the indwelling witness, from a fluent statistical imitation of the same, you'll discover that a whole generation has quietly outsourced the most interior act of the soul.

Thomas More, writing from his cell in the Tower, put the matter more bluntly. "I never intend," he wrote, "God being my good Lord, to pin my soul to another man's back." That's a line that should be printed inside every Catholic classroom in England and Wales this year. Because the pastoral task in front of this Conference is to prevent an entire English generation from pinning its soul to the back of a machine.

With all of that in view, four things these systems simply cannot do.

They cannot know you. They have no interior life.

They cannot love you. Love is the willing of the good of the other. A machine has no will.

They cannot forgive you. Only the priest, standing in persona Christi, can do that.

They cannot accompany you. They can only be in the room.

And yet — and this is the pivot I want you to leave this section with — none of that means the machine is necessarily hostile to the life of the Church. A tool that is honestly named is a tool that can be rightly ordered. The machine can surface the Church's memory; it can't deliver grace. The machine can remove obstacles to the encounter; it can't be the encounter. The machine can clear the intellectual debris between a seeker and the altar; it can't stand at the altar. That's the correct pastoral geometry, and if we hold it, we won't be drawn into the false choice the industry is presenting, between worshipping the new god and refusing the new tool.

Let me land this section with one line I want you to carry out of the room.

Your parishioners aren't in danger of believing the machine is God. They're in danger of forgetting that they aren't machines.

V. The Automation of Human Labour and the Crisis of Meaning

The urgency of getting this right isn't abstract. It will be measured, over the next decade, in livelihoods, in marriages, in suicides, and in souls. And that's the reality I want to stand in front of for the next few minutes.

Somewhere in Wolverhampton this morning, a man who drove a lorry for thirty years is opening a letter that explains his cab doesn't need him any more. Somewhere in South London, a paralegal who finished her training in 2024 is realising that the work she trained for is now priced near zero. Somewhere in Leeds, a married couple with a baby on the way are looking at their income and discovering that they can't plan. These aren't abstractions. These are the faces that are about to walk into your confessional lines, your food pantries, your marriage tribunals — in numbers that your dioceses haven't planned for.

Now let me put that in a frame.

For two hundred years, the modern world has answered the question "Who are you?" with a reductive "What do you do?" The Industrial Revolution tied human dignity, quietly but ruthlessly, to economic output. We've lived inside what I call the GDP Era. And we're now, in real time, watching that era end.

Automation is coming for knowledge work through agentic AI. Automation is coming for physical work through embodied AI. There's no sanctuary. For the first time in human history, generating massive economic value won't require massive amounts of human labour.

And this will land hardest on the English economy. A very large share of the United Kingdom's economy sits in services, in finance, in administration, in knowledge work — exactly the layer that this technology eats first. Both the Office for National Statistics and the Bank of England have already reported disproportionate exposure for UK white-collar workers. This isn't a Silicon Valley problem. This is a parish problem in Manchester, Liverpool, Birmingham, London, Cardiff, and a hundred smaller places in between.

What Silicon Valley is offering, in response, is shallow and insufficient. Their answer is Universal Basic Income plus endless digital distraction. Feed the body. Sedate the mind. Sam Altman, CEO of OpenAI that created ChatGPT, has said publicly that AI will drive the cost of labour toward zero. Elon Musk has said that work will become optional. These men aren't fools. They can see where their own technology is going. What they can't see — what no one in Silicon Valley can see, because their ideological tradition doesn't equip them to see it — is that mass displacement isn't primarily an economic crisis. It's a soul crisis.

Viktor Frankl showed it, from the other side of Auschwitz. When the struggle for survival subsides, the struggle for meaning intensifies. He called the place people arrive at, once the basics are taken care of, the existential vacuum. And the UK is already showing the early tremors of that vacuum. Deaths of despair. The collapse of male life expectancy in parts of the industrial North. The fact that the British Government became the first in the world, in 2018, to appoint a Minister for Loneliness — a tacit admission that isolation in this country had become a national concern.

The historian Yuval Noah Harari has given us a phrase to describe the population emerging from this transition. He calls them the useless class. That's his phrase, not mine, and not the Church's. But I want to address the claim buried in the phrase, because the Church's answer to it has to be sharper than it currently is. The danger in front of us is no longer exploitation. It's irrelevance. The system won't crush your people. The system won't need your people.

If the Church's response is to argue that human beings are still economically necessary, we'll lose the argument. The response has to be more radical. The response has to be a refusal of the premise — a refusal, coming from the Conference of English and Welsh Bishops in 2026, of the idea that a person's worth was ever economic in the first place.

There's a political edge to this, and I think it needs to be named in this room, because no one else is going to name it. Historically, the working class's ultimate leverage against the elite was the strike — the threat to withdraw labour. When labour is no longer necessary for production, that leverage vanishes. If the intelligent machines are owned by a small number of corporations, and the masses depend on a universal basic income paid for out of the taxes on those corporations, we haven't built a liberation. We've built a digital feudalism — a society of dependents, not citizens. A Universal Basic Income in that configuration isn't freedom. It's an allowance.

And because the secular world has no spiritual answer to the onset of mass irrelevance, it's offering distraction instead. The Stanford AI Index this year shows a fifty-point gap between experts and the public on whether AI will be good for their jobs. Seventy-three percent of experts expect a positive impact. Only twenty-three percent of the public do. That gap isn't optimism. That gap is fear. And it won't stay passive. It will metastasise, unless something more serious is placed in its way.

What's being placed in its way, currently, is a modern Soma. Immersive entertainment. AI companionship. Synthetic intimacy. An infinite scroll aimed at a finite soul that was made for the Infinite. Augustine saw it sixteen hundred years ago, and his sentence still describes us: "Thou hast made us for Thyself, O Lord, and our heart is restless until it rest in Thee." The restlessness of the twenty-first century can't be sedated by a subscription.

And hear me on one more thing. Pope Leo XIV has already named this challenge at the very highest level. In his first Address to the College of Cardinals on the tenth of May last year — the speech in which he laid out the programme of his pontificate — he said, and I quote him: "In our own day, the Church offers to everyone the treasury of her social teaching in response to another industrial revolution and to developments in the field of artificial intelligence that pose new challenges for the defence of human dignity, justice, and labour."

That isn't a pious generality. That's a Pope, in his programmatic address, naming artificial intelligence explicitly, tying it to the dignity of labour, and situating it directly in the tradition of his namesake Leo XIII and the encyclical Rerum Novarum. The anticipated social encyclical — the one being called Magnifica Humanitas — is expected to be released on the fifteenth of May this year. Which is to say, twenty-two days from today, on the hundred and thirty-fifth anniversary of Rerum Novarum. The bishops in this room will be among the first in the world to read it. The best thing we can do between now and then is to prepare your dioceses to receive it.

Let me say one last sentence before I turn the corner.

The great crisis of our century won't be scarcity. It will be despair. A Universal Basic Income can't fill a hole in the soul.

Now the turn.

So the question I want to leave with you — the question that the second half of your day, and the second half of my argument, depends on — is this. What does the Church become, when the market no longer requires human labour?

VI. The Church as the Ark for a Post-Work World

Here's what I want to say to this Conference, as directly as I can.

The collapse of the GDP Era isn't a funeral. It's an unveiling. It's the greatest opportunity for evangelisation since the fall of the Roman Empire.

For two hundred years the market has competed with the altar for the heart of man. It demanded his time, his energy, his anxiety, his ambition. It promised him salvation through productivity. And it left the Church the scraps of Sunday morning. That competition is ending. The machine is coming to take the toil. It's coming to take the anxiety of survival. And it's handing back to humanity the one asset we've been too busy to steward. It's handing back time.

I asked you at the start to remember the Easter surge. I want you to remember it again now, because it's already the first evidence of what I'm about to describe. More than a quarter again as many adults received in a single year. Nearly eight hundred of them in Westminster. Southwark's highest count since 2011, half of them aged thirty-five and under, with that striking and specific return of young men. That isn't a marketing success. That isn't a programme working. That's a generation that has been offered everything the digital world can manufacture, arriving — in the silence of the Easter Vigil — and asking for something the digital world can't produce.

The Church holds — and has held for two thousand years — a definition of the human person that no market, no state, and no machine has been able to replace. We're not thinking machines. We're sub-creators, made in the image and likeness of God, willed, as Gaudium et Spes puts it, for our own sake. When the GDP Era ends, the world will desperately need this definition. The Church must not merely hold it. The Church must offer it — publicly, confidently, in plain English.

Now — a distinction to carry back to your dioceses. I want to propose it as a unit of pastoral vocabulary for the next ten years. Toil, and work.

John Paul II taught it in Laborem Exercens. Toil is servile labour. Sweat of the brow. The after-effect of the Fall. Technology can and should lift toil. Work, in the deeper sense — what the Greeks called poiesis — is creative participation in God's own creative act. Gardening Eden. Raising a child. Writing a poem. Tending the sick. No machine can do this, not because the machine is incapable, but because it has no soul.

The right use of this technology, rightly ordered, isn't the end of work. It's the end of toil. It's the first chance, at scale, in human history, for men and women to work for love rather than for survival.

And the Pope has already told the young generation what to do with that possibility. Pope Leo XIV, speaking at the Jubilee of the World of Education on the thirtieth of October last year, in the Paul VI Audience Hall, said this. Listen carefully to the verbs. "Do not let the algorithm write your story. Be the authors. Use technology wisely, but do not let technology use you." That's the charge. It was spoken to the next generation. It was meant also for the shepherds who will form them.

Now — four practical shifts that flow from this. I offer them as the four handles of the Ark, and they will set up the afternoon.

The first shift is democratising the cognitive core. The deepest wisdom in human history has been locked away — in libraries, in Latin, in dense academic books, in archives that most of your parents and grandparents were never going to read. Rightly ordered Catholic AI can turn that static library into kinetic energy that a father can use at his own dinner table with his thirteen-year-old. This afternoon I'll show you, very practically, what that looks like.

The second shift is reframing the Liturgy as the anti-algorithm. Josef Pieper, writing in the rubble of post-war Germany, taught that culture flows from the cult. He meant something quite specific. Free time doesn't become leisure — it doesn't become the condition of creativity — unless it is ordered around worship. Otherwise, it degenerates into boredom. In a post-work world, the Mass isn't a competitor to entertainment. It's the only serious answer to it.

The third shift is building tools that are off-ramps, not roundabouts. Engineered to return the person to the parish, not to keep the person on the screen. That's a design principle, not just a pastoral hope, and it can be implemented at the source.

The fourth shift is recovering the human scale of community. The industrial city was built for the GDP Era. As that era ends, we can rediscover the parish not as a branch office, but as the spire at the centre of a human-scaled life. This is the moment for what architects call cathedral thinking. Laying stones for towers we won't see finished.

Now — because I promised you we would come back to this — the warning.

The Glasswing announcement from earlier this month is, in one sense, a technical story. But it's also, in a deeper sense, a pastoral story. It tells us that the digital civilisation in which our people live is more fragile than they know — and that its custody has passed into the hands of a very small number of private, primarily American, corporations. Even the labs themselves are now surprised by what their own models can do.

If the Church doesn't build her own infrastructure, she will be renting intelligence from those corporations. Their values aren't ours. Their incentives aren't ours. And those systems, whether the bishops fully realise it or not, will quietly set the terms on which Catholic teaching is presented in classrooms, in seminaries, in parish websites, in chanceries, and — in time — in catechesis itself.

The principle we need for this already exists. It's called subsidiarity. Leo XIII taught it in Rerum Novarum. The Catechism reaffirms it in paragraph 1883. Apply it to code. Keep the data at the smallest workable level. Build tools that run on your own machines, inside your own walls, aligned to your own creed. This isn't, in the end, a technical decision. It's a strategic one. And it's a decision that only the bishops in this room can make, for their own dioceses and their own people.

I'm not asking this Conference to become a technology company. I'm asking this Conference to refuse to let a technology company become its Church.

This afternoon, I'll put practical tools in your hands. I'll show you what a rightly ordered Catholic AI looks like inside a marriage tribunal, a parish office, a secondary school, and a domestic home — so that when you leave Palazzola and go back to your dioceses next week, you'll leave not just with a map, but with something to build.

Closing — Do Not Be Afraid

Let me close where we began.

We began with memory. With the generation of English priests who walked out of their College in Rome in 1579, knowing what waited for them at home — and who built anyway. They faced an uprootedness more violent than ours. They responded not by shrinking, but by going deeper.

Four sentences, then, before I stop.

We follow a God who did not stay in the cloud of heaven. He took on flesh, and walked among us, and let us nail Him to a tree.

We follow a God who did not send an algorithm. He sent His Son.

We follow a God who did not optimise. He loved.

We follow a God who did not solve the problem of human suffering by abolishing suffering, but by entering it.

Here's the charge I want to put on your desk.

We'll use the cloud, but we won't live in it. We'll use artificial intelligence to protect real wisdom. We'll use the speed of the processor to defend the slowness of prayer. We'll use the efficiency of the machine to buy back the time we need for charity.

The Holy Father has told us what to tell the young. And I want to hand you his line once more, because it's the seal on everything I've tried to say this morning. Pope Leo XIV, speaking to a generation that is about to inherit this machine, said — "Do not let the algorithm write your story. Be the authors. Use technology wisely, but do not let technology use you."

That's the charge you carry back to your dioceses.

And now, finally, one of the oldest and most repeated commands in Scripture. The line by which a Polish Pope opened the pontificate that ended an empire.

Do not be afraid.

Do not be afraid of this technology. It cannot carry the Cross. It cannot offer the Eucharist. It cannot love your people. But you can. And the reason we've been given these tools — the reason this technology has arrived in our hour, and not in some other — is precisely so that you can do that more fully, not less.

The Easter Vigil has told us what your people are hungry for. The next ten years will decide whether the Church has the courage, and the infrastructure, and the confidence in her own tradition, to feed them.

Let the machines carry the weight of the world.

Let us, at last, carry one another.

Thank you.