This is the transcript of a speech by Matthew Harvey Sanders, CEO of Longbeard, delivered on Tuesday, December 2, 2025, at the Pontifical Athenaeum Regina Apostolorum in Rome. The speech was given during the International Congress Artificial Intelligence as a Catalyst for Contemplation on Human Nature and the Image of God, held in collaboration with the Jaki Stanley Society.

Speech Introduction: The Eve of Contemplation

Your Eminences, Excellencies, Reverend Fathers, distinguished scholars, and friends.

It is a profound honor to be with you this evening at the Pontifical Athenaeum Regina Apostolorum. We are here to inaugurate a vital conversation organized by the Stanley Jaki Society—a society dedicated to a man who understood that the study of the physical world inevitably leads us back to the Creator.

We are gathered tonight on the eve of a significant program. Tomorrow, you will hear from brilliant minds—theologians, philosophers, and ethicists—who will dissect the “Algorithms of Ethics,” explore “Mind as Machine,” and debate the “Violation of Cognitive Freedom”. 

They will grapple with the why and the what of this technological revolution. They will provide the anthropological and moral framework we so desperately need.

But tonight, before we dive into those deep waters tomorrow, I want to offer you a different perspective. I am not a theologian. I am not a philosopher. I am a builder.

My job, and the mission of my crew at Longbeard, is to take the high ideals you will discuss tomorrow—the dignity of the human person, the demands of the common good, the nature of the soul—and translate them into code.

We stand today at a “Digital Rubicon”. We are transitioning from an age of information to an age of automated reasoning. The question before us is not whether we should cross this river—we are already in the water. The question is: who will write the code that governs the other side?

Will it be a code of radical utility and profit maximization? Or will it be a code rooted in the Logos, ordered toward the true flourishing of humanity?

Tonight, I want to share with you how we are attempting to build the latter. I want to talk about “Catholic AI”—not as a marketing slogan, but as a technical reality. And I want to share a vision for “Sovereign AI” that restores power to the individual, turning our technology from a master back into a servant.

Part 1: The Anatomy of an LLM

To understand why a “Catholic-aligned” intelligence is necessary, we must first demystify what an AI actually is.

Building a Large Language Model (LLM) is not magic. It is a recipe that requires three specific ingredients.

First, you need Computers. This is the raw horsepower—the warehouses full of GPUs that process billions of operations per second. Second, you need Architecture. This is the software structure, the neural networks that mimic the connectivity of the human brain.

But the third ingredient is the most critical: Data.

An AI model is only as good as the diet it is fed. 

The secular models that dominate our headlines today—the ones built in Silicon Valley—have been fed the entire internet. They have ingested the collective output of humanity: the profound and the profane. They have read Shakespeare and Scripture, yes, but they have also consumed every Reddit thread, every conspiracy theory, and every expression of moral relativism available online.

When you ask those models a question about the nature of the human person, or the morality of an action, they do not give you Truth. They give you the statistical average of the internet. They give you the consensus of the crowd.

We realized early on that if we wanted an AI that could serve as a true “catalyst for contemplation on human nature,” we could not simply put a “Catholic wrapper” around a secular brain. We had to change the diet.

Part 2: The Foundation – Digitizing the Patrimony

This realization birthed our foundational mission: the digitization of the Church’s patrimony.

We looked around and saw a tragic irony. The Catholic Church is the oldest institution in the West, the guardian of 2,000 years of intellectual treasure. But where was this data?

It was in large part locked away. It was sitting on shelves in monasteries, in the dusty basements of universities, and in archives here in Rome. It was trapped in analog formats—invisible to the digital eyes of the future.

If we did not digitize this wisdom, the AI models of tomorrow would simply not know it existed. Augustine, Aquinas, the Desert Fathers—they would be reduced to statistical noise.

So, we built the Alexandria Digitization Hub.

Right here in Rome, in collaboration with the Pontifical Gregorian University and piloting with the Pontifical Oriental Institute, we have deployed state-of-the-art robotic scanners. These machines are tireless. They gently turn the pages of ancient manuscripts and rare books, converting them into digital text.

But we don’t just scan them as images. We feed them into Vulgate AI, our processing engine. Vulgate AI uses advanced optical character recognition to turn those images into searchable data, and then it goes further—it uses semantic analysis to understand the concepts within the text.

We are effectively expanding the “Catholic dataset.” We are ensuring that the entire intellectual tradition of the Church—her philosophy, her theology, her social teaching—is available to train the next generation of intelligence.

Part 3: The Application – Magisterium AI

The first fruit of this labor is a tool many of you may know: Magisterium AI.

Magisterium AI is what we call a compound AI system. But I prefer to think of it as a digital librarian.

Unlike a standard chatbot that might hallucinate or make things up, Magisterium AI is disciplined. When you ask it a question, it does not scour the open internet. It consults a specific, curated database of over 29,000 magisterial and theological documents as well as specialized contextualization tools. It reads the Encyclicals, the Decrees of Councils, the Code of Canon Law.

And importantly, it cites its sources.

Because it anchors every response in these primary texts, when you use Magisterium AI, you are not really interacting with a machine. You are interacting with the writings of the Popes, the Church Fathers, and the Saints. The AI is simply the steward that sources the relevant material, distills it, and places their wisdom before you.

The global hunger for this kind of trustworthy technology has been staggering. Today, Magisterium AI is the number one answer engine for the Catholic faith in the world. It is used in over 165 countries and communicates in more than 50 languages.

But we didn’t just build it as a standalone product; we built it as infrastructure for the entire Church. We developed an API that allows other organizations to build faithful applications on top of our engine. A prime example of this is the Hallow app. Hallow uses Magisterium AI to power its chat feature, bringing faithful answers to millions of prayerful users. 

We are effectively providing the “theological brain” for the wider Catholic digital ecosystem.

In Laborem Exercens, St. John Paul II taught us that work should elevate the human person, not degrade him. Too often, academic research is drudgery—it is the “toil” of the archive. Magisterium AI takes on that toil so that the scholar can focus on the “fruit” of insight.

However, Magisterium AI is primarily a research tool. To truly impact the culture and protect human agency in the coming years, we need to go further. We need to move from “research assistants” to “personal agents.”

Part 4: The Vision – Ephrem and Sovereign AI

This brings me to the heart of what I wish to propose to you tonight. The future of AI cannot just be about massive, monolithic brains in the cloud owned by a few global corporations. That path leads to a dangerous concentration of power and a potential “technocratic oligarchy”.

We are charting a different path. It is called Sovereign AI.

We are developing Ephrem, the world’s first Catholic-aligned SLM—a Small Language Model.

In the tech world, “Small” does not mean “lesser.” It means specialized, efficient, and runnable on personal hardware. The vision for Ephrem is that it will not live on a server farm in Virginia or California. It will live with you. It will run on your personal compute—your laptop, or a dedicated device in your home.

Think of the character Jarvis AI from the Iron Man films. Jarvis was not a search engine; he was a personal agent. He knew Tony Stark, he protected him, he served his specific needs. 

We want Ephrem to be that for the Catholic world.

Imagine a system that aggregates all your personal data—your calendar, your emails, your health data, your financial records—but keeps it all locally, in your home. You own the data. You control the intelligence.

This approach is driven by the vital Catholic principle of Subsidiarity.

Just as the Church teaches that governance should function at the most local level possible—starting with the family—we believe our digital infrastructure should be organized the same way. We should not surrender the intimate details of our lives to a centralized corporate authority. By moving the intelligence to your own home, we are restoring proper order and ensuring that technology serves the family, rather than the family serving the system.

But Ephrem is not just a filing cabinet; it is a gateway and a shield.

There will always be a need for the massive “Super-Intelligence” models in the cloud for heavy tasks—complex physics simulations or global market analysis. But you should not have to expose your soul to those machines to use them.

Ephrem is designed to run inference with those larger models. When you have a complex request, Ephrem takes it, anonymizes it—stripping away your identity—sends the query to the cloud, retrieves the answer, and brings it back to you.

Crucially, Ephrem acts as an alignment filter. If the secular model returns an answer that is biased, utilitarian, or contrary to human dignity, Ephrem—trained on that pristine Catholic dataset we are building in Rome—can flag it. It can say, “This is what the world says, but here is what the Church teaches.”

It empowers the user to engage with the digital world without being consumed by it.

Part 5: The Catholic Advantage

Now, you might ask: “Matthew, why is the Church the one to do this? Why not a big tech company?”

I will tell you something that is becoming increasingly clear in the halls of advanced AI research: the era of believing that “bigger is always better” is coming to an end.

We are witnessing a fundamental breakthrough in how we understand machine intelligence. Leading researchers, such as Andrej Karpathy, are now discussing the concept of a “Cognitive Core.”

For years, the industry believed that to get a machine to reason or understand multiple languages, you had to feed it the entire internet—trillions of parameters of data, much of it noise, spam, and error. But we are discovering that this is not true. We are learning that we can decrease our dependence on massive amounts of data.

It is possible to isolate the specific “core tokens” and algorithmic strategies required for intelligence. We are finding that if you curate the data perfectly—if you feed the model high-density examples of logic, reasoning, and clear language—you can achieve emergent capabilities, like multilingual understanding and complex reasoning, with a fraction of the computing power.

This shift plays directly into the hands of the Church.

We do not need the entire internet to train a model to reason about the Good, the True, and the Beautiful. We need a specific, high-quality dataset. And the Church possesses the most profound “Cognitive Core” in human history.

Our dataset—2,000 years of councils, encyclicals, and theological disputations—is not just vast; it is dense. It is a masterclass in logic and philosophy.

Furthermore, as many engineers in Silicon Valley privately admit, the Church has a unique technical advantage: Radical Consistency.

To train this efficient “Cognitive Core” effectively, the data cannot contradict itself. If you feed a model the shifting values of secular culture, or political platforms that change every four years, the model becomes unstable. It gets confused.

But the Catholic Church possesses a dataset that is, miraculously, consistent. The teaching on the nature of God, the dignity of the human person, and the demands of charity in the Didache of the first century resonates perfectly with the writings of Benedict XVI in the twenty-first.

This combination—the ability to isolate the algorithmic core of reasoning and the possession of a pristine, consistent dataset—allows us to do something the secular world struggles to do. It allows us to train a Small Language Model that is highly intelligent, capable of deep reasoning, and small enough to run in your home, yet robust enough to represent the Faith faithfully.

We have the highest quality dataset in human history, and now, the technology has finally advanced enough to let us use it.

 

Conclusion: The Cost of Silence

Tomorrow, as you listen to the speakers discuss the “Algorithms of Ethics” and “AI and Knowledge,” I ask you to keep this technical reality in mind. But I also ask you to consider the alternative.

We must reflect, with trembling, on the cost of not acting.

What happens if the Church chooses to sleep through this revolution? What happens if we decide that technology is “too worldly” or “too complex” for us to touch?

If we do not digitize our holdings—if we leave the vast majority of our patrimony locked in the physical darkness of archives and basements—we are effectively silencing our own history.

While the secular AI models of the future will certainly know who Augustine and Aquinas were, they will likely know only the surface—the famous quotes, the popular summaries, the “Wikipedia version” of our faith. They will lack the depth, the nuance, and the fullness of the tradition.

Furthermore, we must remember where we are standing. Right here in Rome, within the libraries of the Pontifical universities and religious orders, there are countless manuscripts that have essentially been forgotten. There are works of profound theological, philosophical, and moral insight that have not been read or referenced in centuries.

By failing to digitize them, we are keeping these insights buried. But by digitizing them and adding them to the Catholic dataset, we are allowing for a Great Rediscovery. We are enabling AI to surface a forgotten homily or a lost theological distinction that might be the exact medicine a modern soul needs to hear.

And this brings us to the hardest question of all: How many souls will be lost because of our silence?

How many young men and women, seeking meaning in the “existential vacuum” of the internet, will ask a machine a question about suffering, or love, or God? If we do not provide the data, they will receive an answer synthesized from the moral chaos of the world. They might be led into despair because the specific, beautiful insight that could have saved them was left gathering dust on a shelf in Rome.

If we have the tools to evangelize this new continent—if we have the ability to build a “Cathedral of Truth” that can reach every home and every heart—and we choose not to build it, we will be responsible for that loss. We will have to answer for why we buried our talent in the ground when it was needed most.

We are not just passive observers of this revolution. We are called to be protagonists.

By digitizing our patrimony through the Alexandria Hub, we are preserving the fullness of our memory. By building Magisterium AI, we are equipping the faithful with a shield against confusion. And by building Ephrem, we are handing the power of this technology back to the families and individuals who make up the Body of Christ.

Let us not be the generation that let the light go out in the digital age. Let us instead be the builders who ensured that when the world looks into the digital mirror of the future, it does not see a machine, but rather, it sees a reflection that points back to the true Image of God.

Thank you, and I look forward to learning from the rich discussions that will be shared over the next two days.