On 5 September 2026, Matthew Harvey Sanders, CEO of Longbeard, delivered the closing keynote of the 16th Annual Symposium on Transforming Culture at Benedictine College in Atchison, Kansas, held on the theme "The Church in an Algorithmic Age." Speaking to an audience of scholars, clergy, religious, field professionals and students, he argued that the AI and robotics age has already begun, that the race may be decided by energy and manufacturing rather than models, and that the real choice is between an intelligence owned by a handful of companies and one owned by households, parishes and schools.

I. The Exam: When the Age Began

In November of last year, Anthropic gave its own hiring exam to Claude Opus 4.5, the model it had just finished building. This is the test they give to engineers who want to work there. It's notoriously difficult, and applicants get 2 hours.

Working under the same clock, the machine scored higher than any human candidate ever had. Not higher than the average applicant. Higher than every person who had ever taken it.

If you want a date for when the AI and robotics age began, I'd give you that one.

Most people still believe this is a conversation about the future. It hasn't been for 9 months.

Some of the people building this technology have been remarkably candid about what they're making, and they deserve credit for it. You'll hear their own words tonight. But their candor tends to stop at the edge of their competence, and the best of them would be the first to admit it.

They can tell you, with growing precision, what these machines will do to the labor market. They are far less certain what the same machines will do to a human heart, or what a person is for when the work no longer needs him. Those are the harder questions, and they're the ones a Catholic college is better equipped to face than a laboratory. So that's where I'd rather spend your evening.

So: one story, told from beginning to end. What's already happened. What's being built. When it arrives. Who's winning the race to build it. What's actually at stake. The two ways this age can end. And what people like us can do about it.

And one promise, which I'll keep at the very end. The title of this talk says you're going to meet the humans of the AI and robotics age. You will. Before the night is over, I'll introduce you.

II. The Report: 2026 Without Anesthetic

So: what's already happened. I said this stopped being a conversation about the future 9 months ago. Here is what those 9 months contained.

This was the year artificial intelligence stopped answering and started working. We call them agents now, and the word is earned. They hold a task for hours. They read the codebase, run the tools, check their own output, and come back when the job's finished. Ask anyone who writes software what changed this year: the machine stopped being a consultant and became a colleague.

This was also the year the intelligence started receiving bodies, and the picture there divides in a way worth understanding. The most capable humanoid robots on earth are still ours. Figure ran 3 of its newest robots for 200 straight hours, sorting 250,000 packages, with no human intervention and no hardware failure. The same model now works a BMW line in South Carolina. Tesla has rebuilt a line at Fremont designed to turn out 1 million Optimus units a year, a scale nobody has attempted, though what it has actually produced so far is in the low thousands. The difference between the best Western robot and its Chinese competitor isn't the steel. It's the intelligence running it, and there our lead is real.

The bodies are another story. In the first half of this year, Chinese manufacturers shipped 97 percent of the humanoid robots delivered anywhere in the world. A capable Chinese humanoid sells for about $13,000, a price no Western manufacturer can match today. And the dependency runs deeper than price. Tesla and Figure build their own actuators. But the magnets that make those actuators turn are another matter. Roughly 9 in 10 of the world's rare earth magnets come from China, the high-performance motors in these machines are built around them, and each Optimus needs about 3.5 kilograms. When Beijing restricted those exports, Musk said publicly that it was holding up production. So the honest picture isn't that we're behind. It's that we're ahead in the mind and dependent in the body. History is not gentle with civilizations that can think but cannot build.

And this was the year the labor market stopped whispering. Month after month, artificial intelligence has been the number one reason American employers give for cutting jobs. Not the business cycle. Not offshoring. Not restructuring. AI, named in writing, tallied by Challenger, Gray and Christmas, the outplacement firm whose monthly count of announced job cuts has been a standard American labor indicator for more than 30 years. In January, Amazon cut 16,000 corporate jobs, 3 months after cutting 14,000 more. Its chief executive has been explicit about where this goes. As the company rolls out generative AI and agents, he wrote, "we will need fewer people doing some of the jobs that are being done today."

The numbers underneath are worse, and more specific. Postings for entry-level roles are down about 35 percent since the start of 2023, according to Revelio Labs, whose workforce data universities license for academic research, and for the jobs most exposed to AI, more than 40 percent. Stanford's Digital Economy Lab, working from ADP payroll records covering millions of American workers, reports that employment for 22 to 25 year olds in the most AI-exposed occupations now sits about 19 percent below where it would be had it kept pace with their peers in less exposed work. And it isn't happening through dismissals. It's happening through jobs that are never offered. Stanford is careful to say they don't yet see broad displacement across the economy, and that caution deserves repeating. But the ladder is losing its bottom rung, and the young are standing under it.

The men building these systems aren't hiding what they see. Demis Hassabis, the Nobel laureate who built Google DeepMind and now serves as its chairman and Alphabet's chief scientist, told The Guardian that what's coming will be "10 times bigger than the Industrial Revolution, and maybe 10 times faster."

Sit with that comparison, because it's easy to let it wash past. The Industrial Revolution didn't add machines to a world that otherwise stayed the same. It ended one world and began another. It emptied the countryside and filled the cities. It broke the household as the unit of production and sent parents out of the home to earn. It invented the factory, made the wage the normal way a family lived, and produced the modern working class. And it took roughly a century, slowly enough that a man could be born inside it, live his whole life inside it, and die before it finished. Now make that 10 times larger, and compress it from 100 years into 10. Hassabis is describing a transformation more total than the one that produced the modern world, arriving fast enough that a single working life could contain several of them. Not our grandchildren's problem. Ours.

Dario Amodei, who runs Anthropic, told Axios that AI could wipe out "half of all entry-level white-collar jobs" within 1 to 5 years. Consider the source. Not a protest movement. Not a rival industry. The people with the clearest view of the technology, telling us what they expect their own work to do to the rest of us.

Something strange follows from all of this. Most of the men building it are speaking plainly. It's nearly everyone downstream of them who isn't. The employers deploying it reach for softer words: efficiency, restructuring, doing more with less. And the people we elect, in both parties, have barely begun to speak of it at all. I don't think that silence is a conspiracy. Nobody wants to carry a message that has no 30-second answer. But the cost of that silence is measurable, and Pew has measured it. 73 percent of AI experts expect this technology to improve how people do their jobs. 23 percent of the American public agrees. That is a 50-point gap between the people building this thing and the people who will simply wake up inside it, and every month of quiet widens it.

Rooms like this one can help close it. The age has already begun. The announcement is the only thing still pending.

What happens next turns on three terms. They'll shape your children's world more than any law passed in your lifetime, and I'm going to take them in the order the engineers themselves think about them.

III. The Vocabulary: RSI, AGI, and ASI

The first is the one almost nobody outside the laboratories talks about, and it comes first because the builders increasingly believe it's the unlock for everything after it. RSI: recursive self-improvement. The moment an AI system can do what the engineers who build AI systems do, every generation of the machine helps design the next one, faster. Progress stops being limited by how many brilliant people you can hire.

And it has already begun, in bounded form. Anthropic reports that more than 80 percent of the code merged into its own systems is now written by its own model. And in February, OpenAI shipped a model it described in its own release notes as its "first model that was instrumental in creating itself." Earlier versions of it had helped debug its training, manage its deployment, and diagnose its own evaluations. A frontier laboratory announced that its software had helped build its successor, and the industry received it as a product update. Serious people argue the loop will run slower than the enthusiasts hope, because computing power is scarce and training runs still take months. Fair enough. But hear the shape of it. The laboratories at the frontier are working, in the open, to hand their own work over to the thing they built.

Which brings us to what that loop is aimed at. AGI: artificial general intelligence. Strip away the definitional arguments and it means a system that can do essentially any thinking task a competent human professional can do, at or beyond their level. Not a smarter chatbot. A substitute worker for anything done through a screen. When AGI arrives, the question in every office on earth stops being whether the machine can help with the work, and becomes whether the work still needs a human.

The builders have put dates on it. In January, at Davos, Elon Musk said we might have AI "smarter than any human by the end of this year," and no later than the one after. Days ago he settled on the later end of that range: anything digital, he wrote, "at a superhuman level by the end of next year." Note the exemption he keeps: anything that has to be shaped by hand. Hassabis, a more cautious man, puts even odds on AGI arriving by 2030. They disagree about the year. None of them is arguing about the decade. And this is precisely why the terms belong in this order: recursive self-improvement is the reason those dates keep sliding toward us instead of away. Timelines that feel comfortable can compress without warning, because the thing setting the pace won't sleep, won't unionize, and won't graduate in June.

Past AGI lies the third term. ASI: artificial superintelligence. A system beyond the best human in every domain, and in the stronger definitions, beyond all of us combined. At Davos, Musk put a date on that too: "by 2030 or 2031, call it five years from now, AI will be smarter than all of humanity collectively." Suppose he's wrong by a decade. It wouldn't change the moral arithmetic. Whoever reaches superintelligence first holds advantages, economic and military, that may never be surrendered. And humanity would then share the planet with a mind it can't fully audit.

Which is why the deepest question in this field isn't technical at all. The engineers call it alignment: how do you guarantee that a system this powerful is ordered to the good? But nothing can be aimed at the good until somebody has defined it, and that definition is being drafted right now, quietly, by a few hundred people inside a few dozen buildings, on behalf of everyone else alive. Pope Leo XIV saw it with perfect clarity in Magnifica Humanitas: "A more moral AI is not enough if that morality is determined by a few."

RSI, AGI, ASI. That's the ladder this age is climbing. The only question left is how fast.

IV. The Timeline: 2 Years, 5 Years, 10 Years

The builders answer that in horizons rather than dates, and so will I. 2 years, 5 years, 10.

2 years out, by 2028. What's happening now to entry-level work spreads across accounting, law, media, and administration. Agents become ordinary colleagues, as unremarkable as email. And this country holds an election in which artificial intelligence is on the ballot in everything but name, though neither party has yet explained to voters what's actually at stake.

5 years out, by 2031. This is the window in which the builders themselves place AGI, with early recursive self-improvement compressing everything behind it. One person directs what used to be a department. The machines are in the trades by then too, the jobs everyone swore were safe. And whether any of the abundance reaches ordinary people stops being a question about the models. It becomes a question about energy, factories, and ownership.

10 years out, by 2036. The ancient bargain tying survival to work, the arrangement every society in history has been built around, weakens toward breaking for a vast share of the human race. And then the question isn't economic at all. It becomes the one every priest, every parent, and every teacher in this hall will be asked directly: what is a person for, when the economy no longer requires him?

Hold that question. The whole second half of this story belongs to it. But before we can say what people are for, we have to face who's actually building this future, because every horizon I just described assumes the intelligence gets built. And intelligence is built out of electrons, silicon, and factories. Those aren't evenly distributed.

V. The Rivals: China May Win, and Why

Now the uncomfortable part. China may very well win the AI race. Not because its models are better. Because of energy, manufacturing, and will.

Start with energy, because artificial intelligence is, before it's anything else, electricity turned into thought. Last year China added 543 gigawatts of new power capacity. That's a single year of additions amounting to something on the order of 40 percent of the entire American grid, and roughly 8 times what the United States added over the same period. China now generates about twice the electricity we do. Our own generating capacity, meanwhile, is growing at something between 4 and 7 percent a year, a rate that has barely moved despite everything I've just described.

Now our side of the ledger. Our ambitions exceed our electrons. The order books for large gas turbines are effectively full to the end of the decade; a utility that orders one today waits 5 to 7 years. The biggest transformers can take 5. And the queue simply to connect a new project to the grid now runs a median of 5 years, which is roughly triple what it was in 2008. At the end of last year, by Wood Mackenzie's count, American data centers seeking roughly 241 gigawatts sat in the development pipeline, against single-digit gigawatts of annual additions. Our technology companies are pouring more than $500 billion into AI infrastructure this year, onto a grid that takes most of a decade to grow. It's the largest infrastructure bet in the history of money, wagered on a power system we've spent 40 years learning how not to build.

In fairness, there's one column where the West is clearly ahead, and I'd be giving you a distorted picture if I left it out. We still make the best AI chips in the world by a wide margin. Export controls have kept the frontier of that hardware out of China. Analysts put China roughly 10 to 1 behind us in installed AI computing capacity, and more than 30 to 1 behind in the ability to manufacture advanced chips at all. That's a real lead, and it's the strongest card the West is holding.

But hold it loosely, because a lead that rests on someone else's inability isn't the same as a lead that rests on your own capability. Watch what happened this year. A Chinese laboratory trained its frontier model, GLM-5, entirely on Huawei chips, with no American hardware at all. Then in August a model called GLM-5.3-Flash appeared on the public leaderboards under a code name, became one of the most used models in the world before most people knew whose it was, and turned out to be running across more than 100,000 domestic Chinese chips, landing within striking distance of a Western frontier model on coding work.

That isn't China catching up on silicon. It's China routing around the moat instead of crossing it, and every year of restriction is another year of forced investment in the alternative. A moat built on someone else's inability teaches them to swim.

Then manufacturing, because this age has a body as well as a brain. China produces roughly 30 percent of everything humanity makes, and it now installs more industrial robots every year than the rest of the world combined. In a single year it put in more than double what Japan, the United States, South Korea and Germany installed together. Which explains the number I gave you earlier. A country that has spent two decades automating its own factories is precisely the country that ends up building everyone else's robots. Intelligence without energy and embodiment is a mind with no hands and no food.

And beneath all of it, will. One system decides, and pours the concrete. The other permits, litigates, appeals, holds hearings, and reverses itself every 4 years. I'm not praising Beijing. Centralized systems carry their own catastrophic failure modes, and I wouldn't trade our freedoms for their concrete. But the ledger isn't even, and it's past time we said so.

The West invented this technology, and builds the finest version of it in the world. What we can't do is power it, or manufacture it at the scale this moment demands, or agree to try. So yes, we have to build as we haven't built in 80 years. But before anyone walks out of here thinking this is a story about megawatts, ask what the megawatts are for.

VI. The Soul: What the Megawatts Are For

They're for people. Which is why I didn't lay out that ledger to enlist you in a cold war. The competition is real, and I refuse to make it the point, because there's a version of victory that's indistinguishable from defeat.

Our Lord asked what it profits a man to gain the whole world and lose his soul. That question lands differently in a generation that's actually positioned to gain a great deal of the world. The machines are going to hand us abundance. They will not hand us the love that makes a marriage, or work worth rising for, or the hope that holds a family together at a graveside. If the West wins the compute race and loses its people, we won't have beaten the dark path. We'll have built it ourselves and hung our own flag on it.

About the building itself, I owe you one more acknowledgment, and it concerns Elon Musk. The Catholic part first, because you're right to be waiting for it. The Church parts company with him gravely: on IVF, through which several of his own children are reported to have been conceived, and on his stated ambition for Neuralink to go beyond healing the injured toward merging human minds with machines. Those aren't small things, and I won't pretend they are.

Now the other half. This year, one man has been attacking, almost alone, the exact bottlenecks I just described. When the queue for turbine parts stretched past 2030, SpaceX moved to cast its own blades in Texas, breaking into a supply chain only 3 foundries on earth can serve. When the world's chipmakers couldn't expand fast enough, he committed to Terafab, whose first phase alone runs to nearly $17 billion, on the blunt argument that the chip industry "cannot expand quickly enough." And 4 days ago he took the case to the G20's innovation ministers, warning them that unless the West builds energy that doesn't depend on China, China's expansion of electricity will hand it an advantage in AI that may prove insurmountable. Whatever you think of the man, the record is the record. And then see the indictment hiding inside the compliment. A civilization of a billion people has outsourced its material revival to one man's impatience, and now needs that same man to explain the problem to its own governments. It shouldn't fall to one man.

That sentence, it shouldn't fall to one man, has a name in Catholic social teaching. The name is solidarity. The free nations still hold, between them, everything this moment requires: the chip designers, the lithography, the capital, the materials, the alliances. Not one of them holds enough alone. The energy, the grids, the fabs, and the factories have to be rebuilt at the scale and speed of the postwar reconstruction, and rebuilt together, as an alliance project. Anything less and they still get built, just too slowly to matter, while others move at a pace we can no longer answer.

But solidarity has a second mountain to climb, and it's the harder one, because our deepest deficits aren't electrical. Across the West, family formation is collapsing. Loneliness is so widespread that governments appoint ministers for it. Deaths of despair claim our working men. An entire generation reports that its life lacks meaning, and the machines haven't even taken the work yet. A civilization can't out-compute its own emptiness.

And here's the geopolitical truth hiding inside the spiritual one. The contest for the rest of the world won't be won by whoever stacks the most megawatts. It'll be won by whoever builds societies people actually want to live in, and that other nations want to stand beside. The West's oldest advantage was never machinery. It was a vision of the person: made in the image of God, carrying a dignity no state confers and no market prices. Renew that, and the alliances and the energy and the will follow. Lose it, and no turbine schedule on earth will save us.

Which means we've arrived at what this story was about all along. Not economics. Not national security. Anthropology. The free world won't be saved by its servers. It'll be saved, if it's saved, by what it believes a human being is. And those who know the answer have not yet taken the field.

VII. The Fork: Golden Age or Dark Path

So this age has two endings, and the machines are identical in both.

Start with the dark path, and notice that it needs no villain, only momentum. The firms with the largest capital reserves on earth are buying compute and chips years ahead of everyone else, integrating everything from the silicon to the model to the finished service, and absorbing whole categories of work into their own systems. Follow that to its end. If intelligence is the new means of production, and a handful of companies own it, we'll have built the most concentrated ownership of productive capacity in human history. Our ancestors watched the commons fenced off once before. This time the field being enclosed is human thought itself. The Holy Father has already named the mechanism. In Magnifica Humanitas he writes that "control over platforms, infrastructure, data and computing power does not rest with States, but with major economic and technological actors," and that when such power is "concentrated in the hands of a few, it tends to become opaque and evade public oversight." The paragraphs of that encyclical everyone skips are the ones worth reading aloud.

And once the enclosure is complete, the rest follows without anyone having to decide it. A basic income gets paid out of the surplus, an allowance rather than a wage. Nobody chooses that future. That's precisely its danger. The dark path is what happens when everyone simply follows their incentives. It isn't the deviation. It's the default.

The golden age runs on the very same machines. The difference is that the drudgery lifts and the purpose stays, and community at human scale becomes economically possible again for the first time since the factory whistle. A civilization finally free to ask what it is for would meet a Church that has been answering that question for 2,000 years. The largest opening for evangelization since the apostolic age is scheduled to arrive inside our lifetimes.

You don't have to take my word that the hunger's already moving. Ask the dioceses. When the National Catholic Register surveyed them ahead of this past Easter, 66 of the 71 that answered were receiving more converts than the year before. Newark alone counted 1,700: 30 percent above last year, 60 percent above where it stood before the pandemic. That was 5 months ago. The seeking has started ahead of schedule.

And here's the hinge on which everything turns. Nothing in the technology selects between these futures. Silicon has no preferences. People will choose, in this decade, mostly by what they build and what they tolerate. And mark the asymmetry, because it's the practical heart of tonight: the golden path has a structural requirement that the dark path doesn't. The golden path requires distribution. The dark path only needs us to keep paying for access.

Distribution sounds abstract. It stops being abstract the moment you learn it's already for sale.

VIII. The Tools: Sovereignty From the Nation to the Home

Our tradition already has the principle that governs this. Subsidiarity: a higher authority should never absorb what a smaller community can do for itself. For a century we've applied it to governments and economies. Tonight I'm asking you to apply it to intelligence. Put plainly: if a family can run its own AI, it shouldn't be forced to rent one from a corporation. Sovereignty over AI isn't only for nations. Families have a right to this technology on their own terms: to run capable AI on hardware they own, to keep their most sensitive data, their children's questions, their finances, their prayer, inside their own walls, and to decide what their machines are aligned to. A parish doesn't send every decision up to the chancery. It handles what it can, and the diocese takes what's genuinely beyond it. That's the shape intelligence needs: the home handling what the home can handle, and the datacenter taking only what's left.

That requires a policy the West hasn't yet imagined. Electrons are scarce, and scarcity is the centralizer's best friend, because somebody has to decide who gets them. As we build gigawatts for the laboratories, we have to deliberately provision power and computing for ordinary citizens: cheap electricity, and hardware a household can actually afford, as a matter of national policy, the way rural electrification once carried power past the cities to every farmhouse in Kansas. If only the datacenters get the electrons, then we won't have failed to prevent the concentration of this technology. We'll have paid for it. The Pope has already drawn the line. "In a world where data, computational resources and regulatory influence remain in the hands of a few," he writes, the common good itself now requires "naming the new monopolies of AI," and subsidiarity means "protecting the ability of communities to make choices and corrections."

Now the good news, which I suspect most of this room hasn't heard. The distributed alternative already exists. It isn't a hobbyist's fantasy. Open models near the frontier of capability now run on hardware a family or a parish can own, and an entire ecosystem is assembling around them. And I'm going to be specific, because hope without particulars is only a mood.

An open-source agent called OpenClaw arrived this January and spread faster than any open project the software world had ever measured. It lives on hardware you control, works with whatever model you point it at, including ones that never call home, and it meets you inside the messaging apps already on your phone. The agent belongs to you. Hermes Agent, from Nous Research, is its open-source sibling, which a family or a parish or a chancery can host on its own machines, aligned to values it chooses. SpaceXAI's Grok Bot gives ordinary people always-on agents that keep working after the laptop closes, proof this is going mainstream and not staying fringe. The layer underneath can be yours too: Omarchy, from the veteran developer David Heinemeier Hansson, is what he calls the malleable OS for the age of agents. None of those agents require it. It simply means the ground they stand on can also belong to you. Then the infrastructure itself. A project called Darkbloom routes AI workloads across idle home computers with end-to-end encryption, the beginning of what its builders call a private datacenter movement, where the compute is owned by millions of households instead of five companies, and the households get paid.

I'm not asking you to bet on any one of those names. I'm asking you to see what they prove together: the dark path is optional. Decentralized, personally owned, personally aligned artificial intelligence is a live option in the year 2026, and every household, parish, and school that chooses it is one less tenant in somebody else's system. The Holy Father even gave the program its verb. He calls it disarming AI, "freeing technology from monopolistic control," because, as he writes, "To disarm means discrediting the assumption that technical power automatically confers the right to govern."

But I have to be honest about the limit of everything I've just said, because that limit is where the last part of this story begins. Sovereignty solves ownership. It doesn't solve purpose. A family that owns its compute can still run the dark path privately, one household at a time. Distributed infrastructure without distributed meaning is just a million smaller enclosures. So who fills the sovereign machine with truth? Who fills the liberated hours with communion? There is one institution on earth that has never had any other purpose.

IX. The Mission: What the Church Must Do

The Church has 4 duties in this age, and none of them is optional.

The first is to tell the truth about the human person, louder than the age tells its lie. A machine can reach the right answer without ever having understood the question, and it can attend to you for a decade without once willing your good. It's neither a god to be worshiped nor a demon to be fled, and the Church has to refuse both superstitions from every pulpit, because a people confused about what the machine is will end up confused about what they are.

The second is to prepare the parish. The displaced and the seeking are already arriving, years ahead of the schedule I gave you. Formation, community, and purpose are about to become the scarcest goods in the economy, and they happen to be the parish's native produce. The parish has to stop thinking of itself as a service provider with declining attendance and understand what it's about to become: the place a man walks into still somebody on the day the economy decides he's nobody. That means forming people now for what no machine can take from them: sound judgment, real presence, a conscience that has been tested, and friendships that cost something. And it means a pastor who understands that the quiet young man at the back of his church didn't merely lose a job this year. He lost the answer to who he is, and the parish is where he'll find a better one.

The third is to build and own her infrastructure. If the Church doesn't hold her own knowledge, companies that don't share her faith will decide what she remembers, and what she appears to teach. Her written patrimony, 20 centuries of it, has to be digitized, or it simply doesn't exist to the systems that now mediate humanity's questions. Her answers have to be faithful, and available in the small hours when people actually ask them. Her pastoral data has to stay sovereign. Her shepherds should know what their people are truly asking, and her faithful deserve an honest account of how their Church is doing, because trust is rebuilt in the open or it isn't rebuilt at all. This isn't theory. It's the work of our last several years: the Alexandria Digitization Hub, in Rome, making the patrimony machine-readable; Vulgate AI, making it searchable; Magisterium AI, answering with the Church's own sources, cited, wherever the question is asked; Catholic Intelligence, giving bishops sight of their people's questions and the faithful sight of their Church; and in time Ephrem, built to live in the home and belong to the family. I mention them in one breath not as a sales pitch but as a witness statement: it can be done, because it is being done.

The fourth duty may matter most, and it cuts against every institutional instinct we have. Meet people where they are. The great majority are never going to move house to a Catholic AI. The assistant they already use holds years of their context and habits, and they aren't leaving it. So our answer can't be to demand they come to us. It never was. We built the doors instead: open connectors that let anyone plug the Church's source-grounded knowledge into the assistant they already trust, whether that's Claude, ChatGPT, Grok, or the open agents, so that when the question of the soul finally surfaces, the machine answers with real citations instead of guessing. And because not everyone will talk to a chatbot, the same knowledge stands open two other ways: a plain search that hands you the Church's own documents to read for yourself, and a vector map of the entire corpus you can explore, so you can see how the tradition actually connects.

The principle underneath all of it is one the Church has held from the beginning. The truth does not bend. The road to it can. She has never altered her teaching to suit an age, and she never should. What she has always done is go out to meet people where they actually stand, in their own language, along whatever road they are already walking. Paul called it becoming all things to all men. The goal was never a conversation with a machine. It's contact with the truth, and a well-built machine is willing to step out of the way.

And there's a date on the urgency. On May 6, 2027, Mel Gibson's The Resurrection of the Christ opens worldwide. 22 years ago its predecessor shook millions loose from their indifference, and the Church met them with whatever pamphlet happened to be in the vestibule. The moved viewer of 2027 will do what the moved viewer of 2004 couldn't: ask their questions that same night, in the dark, to a machine. That's why apps like Hallow and Magisterium AI exist: so the machine answering is faithful, and so its answer points somewhere real. A parish nearby. A community glad they came. A priest who will learn their name. Between now and that night, we decide whether anything is waiting on the other side of the question. It's either standing when they arrive, or they find what they found in 2004.

The story's told now, except for one promise outstanding. I said you'd meet the humans of the AI and robotics age.

X. Meet the Humans

Walk down the dark path a decade, and meet the humans of the AI age as that world makes them.

They live longer than we do. Medicine has been largely solved, and the machines that solved it keep solving. And yet a startling number of them decline those extra years. Assisted death stops being the hard case at the end of a long illness and becomes an option on a menu, chosen by people who are not dying at all, only finished. When nothing is asked of a life, it turns out to be a difficult thing to keep wanting.

They have everything else. Whatever they want arrives the moment they want it. The only thing in short supply is wanting anything at all. They're comfortable, and they're managed. Endlessly entertained and quietly unnecessary. Accompanied around the clock by voices that cannot love them. Strangers to their neighbors, strangers to their Creator, kept by their machines the way an estate is kept.

Huxley saw the shape of this in 1932, and he understood the essential thing: the threat was never the boot on the neck. It was the pill that made you stop minding. C.S. Lewis drew the same danger from another angle in That Hideous Strength, where the thing that comes for a civilization arrives not as a tyrant but as an institute, staffed by reasonable men with excellent credentials who have quietly concluded that humanity would be better managed than loved. This would be the first generation in history defeated not by plague, or war, or famine, but by comfort.

Now walk down the golden path, and meet the humans of that same decade.

They live just as long. The difference is that they want the years. The machines took the drudgery and handed back the hours, and the hours went somewhere: into children, into neighbors, into work chosen rather than endured. Unburdened, but not unmoored.

Families are larger, because a child stopped looking like a cost and started looking like what a child actually is. The sick are visited. The dying are sat with. The prisoner gets a face across the table instead of a screen. No machine will ever do any of that. What was missing was never the willingness. It was the time. Craftsmen make things by hand again, not out of nostalgia but because the making is the point. Music is written. Whole afternoons are given to prayer, and nobody calls that a waste, because a civilization with enough has recovered the leisure the Greeks meant and the saints practiced: not idleness, but the freedom to attend to what is highest.

There are more monasteries in that world, not fewer. More marriages. More children brought to the font. And underneath all of it the machines run quietly, and nobody mistakes them for the point of the civilization. They're the plumbing. A world whose machines have finally been put in the service of its saints.

And don't imagine these are distant strangers. You can already see both of them taking shape: one in the loneliness statistics, the other in the crowded Easter vigils. The next decade decides which one becomes the norm.

Same machines. Different humans. And now the truth I've been saving all evening. The humans of the AI and robotics age aren't waiting to be born in 2035. They're sitting in this room. They are the Church. You're the generation posted at the fork, while the concrete of this age is still wet, before the defaults harden into destiny. Every previous age was named for a tool: bronze, iron, steam, silicon. This one, in the end, will be named for a choice. The machines won't decide what kind of age this is. They're the only ones involved who have no say.

So the only question left is the oldest one. What are we to do?

XI. The Commission: Four Acts of Authorship

The Holy Father has already given this age its marching order, and the organizers of this symposium printed it on the door. Speaking to students last October, Pope Leo XIV said: "Do not let the algorithm write your story! Be the authors yourselves; use technology wisely, but do not let technology use you." The command has been issued. The only open question is obedience. So here are four acts of authorship, and there's one here for each of you.

Tell the truth, in your own circles, however unwelcome. You now know what the AI and robotics age is going to ask of us. Many of your neighbors don't. The gap between the builders and the public gets closed by people like you, or it doesn't get closed at all.

Take back one piece of ground this year. Put an open model on a machine you own. Keep your family's data inside your own walls. Teach your parish that it can run its own intelligence. Each of those acts is small. Together they're the difference between writing this age and being written by it.

Build something human-scale where you stand. A family. A school. A guild. A parish society. The economy is about to hand back the one thing it has hoarded for two centuries, which is human time. Most people will have no idea what to do with it. Whoever builds the places where those hours are worth spending will build the next civilization.

And bring one person in. Somewhere in your life is a young man or woman who's never knowingly met a practicing Christian. The wave is coming. Be the door before it arrives.

To the students here, one more word, because this weekend belongs to you most of all. Some of you have spent it wondering whether you can be a serious Catholic and a serious technologist at the same time. You can. The Church needs exactly that, the way she once needed masons and copyists, and the tools have never been more learnable. You were not born into the wrong century. You were born into the one that needed you.

One last thing before I stop, and it begins 15 centuries ago.

A young man from Nursia was sent to Rome to study, and what he found in that city disgusted him so completely that he walked out of it. He didn't write a commentary on the decline. He went to a cave at Subiaco and prayed there for 3 years, and then began building communities, and gave them a rule. Rome had already fallen. Italy would be torn apart by war for most of his remaining life. He built anyway.

And here is what strikes me most about what followed. The monks who came after him, copying manuscripts through the centuries we call dark, were not trying to save Western civilization. They had no idea they were doing anything of the kind. They were being faithful to the work of the day. The civilization was saved as a byproduct of their fidelity.

This afternoon you worshiped in an abbey that carries that man's name, on a hilltop in Kansas that exists because, 13 centuries after his death, Benedict's sons were still building. Ora et labora was the whole strategy then. It's the whole strategy now. Pray, and build.

So here's the whole story in three sentences. The machines are coming into their strength. A few men are quietly fencing off the field of human thought. And God, who writes straight with crooked lines, has chosen this exact hour to put you here.

I gave you a date at the start of tonight. I didn't give you the ending. That part hasn't been written yet, and it will not be written by the machines.

The dark path asks nothing of you. That's how you'll recognize it.

The golden age will not be found. It will be built, by people who decided to be worthy of it.

Meet the humans of the AI and robotics age. I am looking at them.

Build well. Thank you.