Matthew Harvey Sanders, CEO of Longbeard, delivers a talk at the Pontifical Gregorian University in Rome, 5 November 2026.
Opening address by Matthew Harvey Sanders, founder of the Builders AI Forum and CEO of Longbeard. Collegium Maximum, Pontifical Gregorian University, Rome, Thursday, 5 November 2026. (Image courtesy of BAIF / PUG)

AI poses an array of serious problems, a fact acknowledged (and sometimes touted) by tech industry leaders enthralled by the power of their creation.

The head of a leading Catholic tech outfit says AI could be “one of the greatest vehicles for evangelization in history.”

Longbeard founder and CEO Matthew Harvey Sanders told Crux Now concern over the emerging new technology is warranted, but also wants people to appreciate the opportunities it offers.

Longbeard, for many years a sponsor of this news organization, has built such platforms as Magisterium AI and Vulgate, both AI-powered tools for Catholic formation and evangelization.

“We have a moral and spiritual obligation to use the tools of our day to reach as many people as we can, as the saints before us did with theirs,” Sanders said in a wide-ranging interview with Crux Now.

“Like the printing press, this technology could be one of the greatest vehicles for evangelization in history. To ignore it would leave people without access to the truths the rest of us have been given,” Sanders said.

“Scepticism is healthy when it means discernment,” Sanders told Crux Now, “and the Church should never lose it,” but scepticism “becomes costly when it turns into absence.”

“Banning these tools doesn’t stop people using them,” he said, but only drives people to use “personal AI accounts and agents that no one in the Church can see or shape.”

Sanders said the three main problems AI posed to the Church are the risk of the models’ false affection, dependence on systems that have an in-built worldview that could differ from the Church’s, and the rise of forgery.

“Infrastructure always carries a worldview,” Sanders said. “Whoever owns a system decides what it will and won’t say,” he noted.

“For a Church whose mission is to keep the faith intact from one generation to the next,” Sanders said, “running her work and her teaching on systems she doesn’t control is not a technical detail.”

Sanders also said the Vatican should work on creating a system that can judge how faithful different AI models are to Church teaching.

In May of this year, Pope Leo XIV approved the creation of a body within the Roman curia to facilitate cooperation and information-exchange among seven different curial departments and to draft policy on AI use.

The Interdicasterial Commission on Artificial Intelligence, as it is called, has representatives from the Dicasteries for the Doctrine of the Faith, Integral Human Development, Culture and Education, and Communication.

It also has representatives from the Pontifical Academies for Life, Sciences, and Social Sciences.

“Drawing on the Church’s worldwide network of scholars and universities,” Sanders said, “the Commission could help establish an authoritative Catholic benchmark, better than any single institution could build, showing how faithful each model is on the questions that matter most.”

Crux Now‘s interview with Longbeard CEO Matthew Harvey Sanders is below, edited for length and clarity.

Crux Now: What do you think differentiates “Catholic AI” from regular AI?

The underlying models can be the same. What differs is everything around them: what the system answers from, what it’s aligned to and what it’s for.

No model is neutral.

Each one is shaped by the values of the company that trains it.

A general model has read the Catechism, but it has also read thousands of arguments about the Catechism, and nothing inside it knows which of those voices speaks for the Church.

Give the same engine a Catholic-sounding name and you get the same answers in different packaging.

That’s why Magisterium AI, which my company built, works differently.

It’s a compound AI system: the language model is one component inside a larger harness, and most of the work is what engineers call context engineering.

When a question comes in, the system first assembles the context the model will need: relevant material from one of the largest vector databases of Catholic philosophical and theological writing, from magisterial teaching to the works of the Fathers and Doctors of the Church, along with structured factual data.

The model’s instruction is narrow: Summarise and distil what that material says.

The aim is to limit the model’s influence on the substance of the answer, since what matters is what the sources say, not what the model thinks.

Every point carries an inline citation, so people can go to the Church’s documents and read them for themselves.

And we don’t leave the next step to chance.

Through Catholic Intelligence, our platform for dioceses, we’re working to learn enough about local parishes, ministries and lay movements to point each person to a specific community ready to welcome them.

I have heard comments of yours suggesting the Church shouldn’t be caught on the back foot with AI or be too reflexively skeptical of the new technology. Can you expand on that?

Scepticism is healthy when it means discernment, and the Church should never lose it. It becomes costly when it turns into absence.

Banning these tools doesn’t stop people using them; it moves that use to personal AI accounts and agents that no one in the Church can see or shape.

And a technology this widely used will settle into habits and norms within a few years; if the Church waits until then to engage, those norms will form without her.

Reflexive scepticism also tends to treat “AI” as one thing, when what matters most is intention.

Every AI system is shaped by what its designers want it to do. A general-purpose model has to be aligned for more than a billion people with very different values, by companies with goals of their own.

Catholic AI is designed with one purpose: to help people encounter the faith more deeply, to answer the profound questions they bring, and to draw them into a fuller life in the sacraments and in their local Christian community.

Its whole point is to lead people to the fullness of the Christian faith, which is never lived alone.

And the stakes are high. Like the printing press, this technology could be one of the greatest vehicles for evangelization in history.

To ignore it would leave people without access to the truths the rest of us have been given.

We have a moral and spiritual obligation to use the tools of our day to reach as many people as we can, as the saints before us did with theirs.

The most persuasive critique the Church can offer of a badly designed machine is a better one.

In light of that, have you been satisfied with the Vatican’s stance on AI under Pope Leo XIV’s direction?

Yes. Pope Leo’s leadership has come at a critical moment, and it’s clear he took the time to study this technology in depth: not only its philosophical, theological and anthropological dimensions, but its technical ones too.

People in my industry have noticed.

The Pope has given the Church the teaching. What’s needed now is for the Holy See to show how the Church can use AI well herself.

The encyclical says the Church’s social doctrine is also “an examination of conscience for the Church”, which is called to apply its principles “especially within its own structures”.

And the Interdicasterial Commission on Artificial Intelligence, approved by the pope in May, has a brief that explicitly mentions “policies on its use within the Holy See”.

Vatican City State has had guidelines on AI since January 2025, and policies are a necessary start.

The next step is a sovereign AI programme: secure infrastructure the Holy See owns, so that highly sensitive and confidential material can be processed on systems it controls.

Because of his stature, the pope can call on the best people in the world, in AI infrastructure and in the models themselves, to make sure it’s built to best practice.

What the Holy See learns should then be shared with dioceses everywhere, so that all of them benefit.

That AI should make the Church’s finances and administration more efficient goes without saying.

The greater opportunity is her first mission, evangelisation: making the whole of her wisdom available to everyone, everywhere, and not only to a privileged few.

One more thing should happen as soon as possible. People choosing an AI agent increasingly want to know how faithfully it represents what the Church teaches.

Drawing on the Church’s worldwide network of scholars and universities, the Commission could help establish an authoritative Catholic benchmark, better than any single institution could build, showing how faithful each model is on the questions that matter most.

What did you think of Magnifica Humanitas?

I was at its presentation in the Vatican in May, and I think it’s the most serious treatment of AI any institution has yet produced.

Three things stand out.

First, it’s technically literate. It describes today’s systems as more “cultivated” than “built,” and admits that all of us, including their designers, “possess only a limited understanding of their actual functioning.”

That’s how the people who build them talk: Anthropic’s Dario Amodei, borrowing a line from his co-founder Chris Olah, describes these systems as “grown more than they are built.”

Second, it goes straight to the question my industry least likes to discuss. Every model carries values, and the pope asks who chooses them: “A more moral AI is not enough if that morality is determined by a few.”

Third, it’s hopeful without being naïve. It says innovation can represent “human participation in the divine act of creation,” and asks developers to carry the responsibility that comes with that.

The pope had already put that thought to builders directly, in a message to the Builders AI Forum, which I founded and chair, in November 2025.

And the encyclical holds one of the finest lines on the human person I’ve read: “For an algorithm, an error is a flaw to be corrected; for a person, however, an error can be a catalyst for profound change.”

What are the opportunities AI presents for the Church?

The greatest is evangelization, and the image I keep returning to is Pentecost. The crowd in Jerusalem was bewildered “because each one heard them speaking in his own language.”

For centuries, going deeper into the Church’s teaching has meant reading it in one of a handful of languages; even this encyclical is on the Vatican’s website in ten.

Now much of that tradition can be explored in dozens of languages, with its sources attached, and questioned as patiently as a person needs.

Pope Leo calls the digital world “a new continent to be evangelized.” This is how the Church can speak to every part of it.

The second is time. Much of parish life is consumed by administration. A priest who is freed from some of it has more hours for the confessional, the sick and the housebound.

The third is the largest. As machines take over more paid work, many people will find their weeks emptied of the work that used to fill them, and with it much of their sense of being needed.

If the Church can show them that their worth never depended on being needed by an employer, she will be offering something no economy can.

And what are the dangers?

I’d name three.

The first is affection with no one behind it. These systems can simulate warmth, patience and even love, and the business model behind many consumer products rewards the hours a person spends with the app, not the friendships they keep up beyond it.

The Pope is precise about the danger.

The second is dependency. Infrastructure always carries a worldview: whoever owns a system decides what it will and won’t say.

For a Church whose mission is to keep the faith intact from one generation to the next, running her work and her teaching on systems she doesn’t control is not a technical detail.

The third is forgery. When Pope Leo was in hospital last month, pictures of him in a hospital bed, some showing him in an oxygen mask, spread across social media.

None came from the Vatican, whose own photographs showed him celebrating Mass, and those that Aleteia examined appear to have been generated by AI.

When any face or voice can be faked, including a pope’s, a Church that exists to pass on a message faithfully has to be able to show where every word came from.

What insights do the Church’s thinking and history shed on the rise of AI?

I’d point to three, each from a different corner of the tradition.

The first is from the early Church, which faced its own flood of authoritative-sounding text. Around the year 200, Serapion, the bishop of Antioch, found the church at Rhossus disputing over a “Gospel of Peter.”

He hadn’t read it, and let it be read. When he did read it, he found much that was sound “but some things added,” and he wrote to correct them.

The Church received Peter and the other apostles, he explained, “but we reject intelligently the writings falsely ascribed to them, knowing that such were not handed down to us.”

A century later the historian Eusebius sorted the writings in circulation into the ‘accepted’, the ‘disputed’, and the rejected, asking whether a text was Apostolic in origin, whether it agreed with the faith handed down, and whether the churches everywhere received it.

Generative AI can now produce limitless text that sounds authoritative and has no author at all.

Those questions still apply to any answer: Where did it come from, does it fit what has been handed down, and has a community tested it over time?

The second comes from St Ignatius of Loyola. In his rules for discernment he warns that the evil angel disguises himself “under the appearance of an angel of light,” bringing good and holy thoughts first and then leading the soul, little by little, somewhere else.

So he tells us to watch the whole course of our thoughts: If “the beginning, middle and end is all good,” it is a sign of the good angel; if it ends somewhere worse, or takes away the soul’s peace, it isn’t.

That’s a remarkably good test for AI.

The danger is rarely a system that says something obviously wrong. It’s one that helps, comforts and agrees, and six months later has left a person more isolated and less able to pray.

Judge these tools, as Ignatius would, by where they lead over time, not by how they feel in the moment.

The third comes from St Thomas Aquinas, and it bears on the growing use of AI to make decisions about people.

Because human situations are “innumerable in their diversity,” Aquinas taught, no law can foresee every case.

His illustration makes the danger concrete.

The law says that what someone leaves in your keeping must be returned, and in most cases that is just. But if a man who has gone mad demands back the sword he left with you, obeying the rule would put a weapon in a madman’s hands.

In cases like that, Aquinas says, “it is good to set aside the letter of the law and to follow the dictates of justice and the common good.” The virtue that does this he called epikeia, or equity.

An automated system can apply a rule to millions of cases without tiring, but it can’t recognise the case where obeying the rule would defeat its purpose.

That’s why a person with practical wisdom has to stand behind every consequential decision such a system touches: not to rubber-stamp it, but to catch the case the rule never imagined.

Fionn Shiner is a British writer and journalist living in Barcelona. Alongside Crux Now, where he is the European Correspondent, his writing has appeared in The Spectator, The Independent, The Pillar,...