Millions Are Asking AI About God

Matthew Sanders on Magisterium AI, digital evangelization, and where Pope Leo might take the Church in the AI age

In recent years, artificial intelligence has quietly become a new space for spiritual inquiry. Millions of people are now turning to AI not just for information, but for guidance—asking questions about meaning, suffering, morality, and God. What does it mean for the Church when these conversations are no longer happening in parishes or with priests, but through algorithms?

Matthew Sanders, a technologist who has built major digital platforms for the Catholic Church—including projects for the Vatican Observatory and the Vatican’s Office for Migrants and Refugees—has been at the forefront of this shift. His latest initiative, Magisterium AI, seeks to bring the Church’s intellectual and spiritual tradition into direct conversation with users through AI.

In this conversation, we discuss:

• Sanders’ career in technology, design, and spreading the Church’s message

• What people are actually asking AI about faith and God

• The risks of outsourcing spiritual authority to technology

• What Pope Leo might say in his forthcoming encyclical on artificial intelligence

This is a conversation about technology, the Catholic faith, and the unexpected places where spiritual life is beginning to unfold.

Transcript

CNS Note: This transcript was generated with the assistance of AI tools. For precise quotations, CNS recommends referring to the video above.

Intro: Matthew Sanders is a technologist who has created some of the most significant digital products for the Catholic Church in the past decade. He has designed websites for the Vatican Observatory, for the Vatican’s Office for Migrants and Refugees -- advancing a key priority of Pope Francis -- and for several central offices of religious orders in Rome. Most recently, he has tried to bring artificial intelligence into the Church’s efforts to evangelize by creating a chatbot -- Magisterium AI -- built on the advances of ChatGPT. Sanders says that millions of users around the world are now using his product. And while user data is private, he can see the kinds of conversations taking place—highly personal, often deeply spiritual exchanges—and the kind of evolution they seem to encourage in users. In this conversation with Catholic News Service, we talk about Sanders’ career in technology and his service to the Church, the emergence of AI as a tool for evangelization, and where he thinks Pope Leo will take the Church in the digital frontier.

Robert Duncan: Matthew Sanders, thanks for sitting down with Catholic News Service. Tell me a little bit about how you came to be what I think many people know you as, the guru, or at least one of them in the Catholic Church.

Matthew Sanders: I really have no idea. I mean, definitely, I think it’s more down to providence than any kind of intentional decision that I made. More than anything else, I think I was just a guy, like many, who was on a journey. And along that journey, there were times in, you know, becoming Catholic initially and then preparing to be a priest, and then eventually working at an archdiocese. There were just times when I just felt very ill-equipped. And I wish that there were more efficient ways to apprehend the knowledge and impart it in a more efficient way. And I think because it kept coming up, this issue of, I don’t know enough, I need to know more. If I’m going to be more of service to the Church, kept coming up again and again at different intervals in my life. And because I kind of grew up in the technology age, I think I was always kind of waiting for a technology solution to this problem. And so when ChatGPT kind of demonstrated that there might be a way to do this finally, I think I said, this is it. So we got to go all in here. I didn’t, we didn’t really go all in initially, but as soon as it became clear, it might be possible, I said, this is it. We have to make this work. Because I imagine there’s, you know, millions, billions of other people who are probably on a similar journey who could benefit in the same way I did.

Robert Duncan: I mean, I know, though, that your journey with where you are now with Magisterium AI, which we’ll talk about, and your other projects are more recent since ChatGPT. But I mean, you were always a great lover of technology, and you came to Rome, initially building websites. Can you tell me about how you got here?

Matthew Sanders: Yeah, sure. I, when, like I said, I grew up in the technology age and, you know, I fell in love with the church. It became very clear to me that, I mean, the founder of the community that I entered was called, his name was Isaac Hecker, and he’s the founder of the Paulist Fathers. The Paulist Fathers, you know, kind of made it a name for themselves by leveraging the most modern technology at the time to evangelize. You know, they used the printing press and the radio, and eventually went into TV and film production. I always, so I always resonated with the pragmatism of that approach. When eventually I started working with the Archdiocese, initially, we were just trying to leverage technology to make the office run more efficiently. But eventually I moved into the special projects office and ran that, and one of the projects they gave me was, was reaching young people. Basically we were trying to launch student Steubenville Conferences in Toronto. I had never been given a marketing task like that before. So basically put together materials and try to entice the youth to sign up for this conference. And so I, you know, we had to do research. Well, what does it take to, to win on platforms like social media? And of course I was like, wow, this is, this is actually really going to be really hard. Especially at the time that the assets they were giving me to reach young people. I was like, okay, that that’s, those are obviously not anything near what the, what the secular industry is using. So it’s probably not going to be effective on that war for the feed for people’s attention. Anyway, eventually we got the job done, but I realized that the Church has got to leverage technology, especially leverage the web and a much more kind of focused and deliberate way, specifically around the area of evangelization. So we left the archdiocese to found a company and that the company’s kind of mission was to help the Church leverage technology more effectively to, kind of, win.

Robert Duncan: And do you know why that is? Because, I remember, I know what you’re describing because when I came to Rome, there was a lot more opportunity and a lot more need, I think, than there is today. There’s still a lot of need, but it’s changed a lot in 15 years. But even then, like around 2010, if you looked at Protestant and Evangelical outreach online, it was far more sophisticated than the Catholic Church’s. Do you have any idea why that is?

Matthew Sanders: So it’s funny you bring that up. I mean, I, when we were trying to figure out a way to convince people that going to Steubenville, Toronto was going to be worth it, we wanted to look at exemplars. So who had actually, who had done something like this before and did really, really well. We studied LifeTeen, the LifeTeen conferences, the LifeTeen Steubenville conferences were pretty good. They filled stadiums. But you know, the real exemplar, the one who really had really knocked us out of the park and really built a global brand for themselves was Hillsong, which was an Evangelical outfit. And they started, you know, they had a church, but then they, they focused a lot on music and that music basically went global, right? They built like a music empire and that led to a kind of a ministry empire and then these large events where they fill football stadiums and their marketing was top notch. I mean, every piece of collateral that they produced to market what they’re doing, whether it be their music or their conferences was top notch. I mean, just as good as the secular industry. That’s what really impressed me. These are people who recognized how important, if you want to win the feed, if you want people’s attention, you have to invest in making sure that your collateral is engaging. You’re telling your story in a way that people find inspiring. So that’s when I, when I really realized that has to be the bar, like, you know, we have to look at who’s doing, who’s doing the best marketing in the world and then what can we learn from them and try and try to ensure that our own marketing for the Church is up to the same standard, which was, which was a challenge, but I think that’s, that’s one of the reasons why it took, it took, you know, 10 years for the Church to really finally kind of get the message and then be willing to put the resources into actually doing this the right way.

“Who’s doing the best marketing in the world … what can we learn from them … to ensure that our own marketing for the Church is up to the same standard?”

Robert Duncan: It’s expensive.

Matthew Sanders: Right. And I mean, it’s, and a lot of it comes down to tactical expertise and taste, right? So finding people who know how to execute these campaigns, knowing people to just know what a, what a good campaign like looks like, right. You understand where people are at and know what message they need to hear and know how to package it. There’s not a lot of people who know how to do that really, really well. Now there’s lots of people I think who are mission aligned with the church who’d be willing to give their gifts to the Church, but they have to be asked.

Robert Duncan: Can you tell me a few of your early successes in doing this, either with the archdiocese or once you launched Longbeard, I mean, people watching this, they may not even know who you are and what you’ve built. So what are some of the —before we get to Magisterium AI and some of the more AI focused stuff — talking about marketing specifically and making the Church’s message more sort of effective online, what were some of those early projects you had that led you to Rome and then once you got to Rome…

Matthew Sanders: We did a bunch of, we initially started with very small projects, but I remember that one phone call where I was like, it’s really the first time I think where I was like, man, we may have finally started to make a dent outside of the, you know, Toronto, Toronto area was, I got a call from the Monks of Norcia and they wanted help building their website and their, and their beer website. And I was like, the Monks of Norcia, I’ve been listening to them, you know, listening to their chant every morning from my office at the archdiocese for at least I think a year or two. So that was, that was pretty exciting for me. Mainly cause they were kind of heroes of mine, you know, you know, later on we ended up working on the Vatican Observatory, kind of new website as well and helping them.

Robert Duncan: I didn’t know that.

Matthew Sanders: Yeah. So we built their website. I’m, you know, it’s been some time since we’ve, you know, I basically have been custodian of it, but that was a really fun project.

Robert Duncan: And your contact then, that would have been with Brother Guy.

Matthew Sanders: Yeah, that’s right. They were great to work with. We ended up working with the Australian Catholic University, building a product called Ethics Finder, which was another interesting kind of technology project. And then eventually we started consulting here in Rome—more on technological infrastructure—starting with trying to modernize a typical university. So Father David Nazar, who’s the rector of the Pontifical Oriental Institute—and also Canadian, so it’s not just the Italians who practice nepotism—I like to think Father Nazar is just pragmatic. When he inherited the Oriental, he had a number of challenges he was facing. One of them was improving the educational experience, but also finding a way to universalize the institution more—making the knowledge in the library and the expertise of the faculty more accessible. So part of that project was doing an audit: how can we integrate technology more effectively into this building? That was really fun. We installed, I believe, one of the first Wi-Fi mesh networks—certainly among the pontifical universities, maybe even in Italy. We put in robotic cameras in their Aula Magna, their main conference hall, so they could do dynamic live streaming—it was like having a TV crew on site. We set up Google Meet kits in all the classrooms to hybridize them, so anyone—even someone sitting in the back—could raise their hand, ask a question, and people online could hear them. It created a really immersive environment. What was really notable is Father Nazar did all that simply because he thought, “this just makes sense.” And then COVID hit just after we installed everything. So he likes to say they were probably the only pontifical university that never missed a class—they just pivoted straight to online.

Robert Duncan: Do you ever, during that time, feel like you were getting mired in IT-related internal work, where maybe what you originally wanted to do was revamp the Church’s public image online? Or did you see those things as both necessary and important?

Matthew Sanders: I think they were both. I have to say, it did feel at times like I was taking a bit of an off-ramp from what I thought we had built the company to do. But it felt essential. Especially working here in Rome, you realize that if you can’t help decision-makers operate more effectively and efficiently, it’s very difficult to get their time or attention for things like, “Hey, maybe we should be doing a better job on marketing.” And these were immediate problems they were feeling. So I felt like we were trying to build relationships and trust. Instead of going in and telling them, “This is what you need to do—you need to improve your marketing, build a new website,” I just tried to listen and understand what practical problems they were experiencing, and whether technology could help address those. And I found that the more we helped solve those immediate, practical problems, the more open they were to other ideas. A good example is when we started working with the Holy See. At the time, Pope Francis was directly overseeing the migrants and refugees section, and that’s when we started working with Cardinal Czerny—Father Michael Czerny back then. They were setting up a new office and wanted to be as effective as possible. They wanted to leverage the most contemporary technology, and they essentially said, “Let’s just use the best tools that Fortune 500 companies are using”—obviously respecting privacy and security, but otherwise without constraints. So working with them to figure out what that would take, understanding how they operate and what it’s like to run an office with global reach—those lessons became very useful for us later on.

Robert Duncan: And meanwhile, while you were working with the Orientale, I remember watching you sort of take over Rome—at least online. You did the Franciscan website, the Curia of the Franciscan Order. You mentioned the Vatican Observatory. You started working with the Holy See. That must have been exhilarating—to arrive and then just client after client, working your way up to the Holy See.

Matthew Sanders: Yeah, I mean, it was. For a systems guy from Toronto, it was all kind of remarkable. There was no grand strategy where we said years in advance, “We need to land the Holy See as a client.” It wasn’t like that at all. It was really just arriving here, having conversations, hearing problems, and saying, “I think we can figure that out.” And then you solve a few problems, word gets out, and someone else says, “I’ve got a few problems over here.” And that’s really how it all came about. The Angelicum is another example—I loved working with them, building their website. That was a unique experience. The Franciscans, the Benedictines—just getting to understand the different religious orders. They have a lot of problems in common, but also some very unique ones. And as a byproduct of helping them tell their story online, you inevitably get more immersed in their traditions, which I found spiritually very uplifting—not to say there weren’t challenges as well.

Robert Duncan: I want to talk specifically about your consultation client work with the Holy See. How did that come about? And you had been picking up all these ecclesiastical clients, and then you get, let’s say, to the top—to a dicastery level. What had you learned about the way the Church works with technology that maybe was crystallized when you started working with the Holy See?

Matthew Sanders: I remember when we initially started conversations about working, in a case like the Migrants and Refugees section—which later became public that Pope Francis wanted to run it directly because he wanted to see something done there meaningfully. Part of the conversation was: if you were building a dicastery or a section and you had kind of carte blanche on what technologies to use in order to create a modern office, what would you use? And I remember thinking, I mean, I can tell you, but that’s never going to happen. There’s too much red tape—you know how these things go. But they said, no, we’re curious. So I laid it out for them—what I would do, what I’ve seen work, what is kind of industry best practice. And what was shocking to me was that eventually the direction came down: okay, let’s do that. The Vatican is full of great people with amazing intentions, but like any bureaucratic organization, there’s a lot of complexity. Sometimes new ideas take time to work their way through the system. I had presumed something like this would take years. So it was a very unusual experience to have the red tape kind of cut and to be told, just go do it. That was very exhilarating. And what I had hoped was that by demonstrating how these technologies could help an office work more productively, others would see it and say, they’re doing it—so why can’t we? I often joke that in Rome, seeing really is believing when it comes to technology. And I think that’s in large part because many of the people in leadership are academics, like I said, and they haven’t been exposed to a lot of these technologies or how they solve problems and make things more efficient. So they don’t really have an imagination for it.

If you tell them this could save 50% of your time, it’s hard for them to understand what that means—they need to see it. And I’ve found so far that pilots are usually the most effective way.

Robert Duncan: I would imagine that getting the Holy See—the Vatican—as a client is a hard get in principle. It sounds like it, at least to a lot of people, maybe in your industry, and possibly because a lot of people there don’t speak your language. It takes a lot of translation on your part to do what you did. So I’m wondering if you have any advice—sort of an aside—that you would give to people working with clients in general, maybe something you learned from working with the Holy See.

Matthew Sanders: I think it really comes down to this. When you have this instinct that there’s a problem you need to solve, and you find others who believe the same thing, you kind of want to go on a crusade. You want to share it, make people aware of the problem, get them to agree, and say, “Give me some money and I’ll help you solve it.” The problem is, not everyone’s there. And trying to convince people there’s a problem they’re not ready to acknowledge—or feel like they can’t because they have too many others—that can be a grind. What I found is the best strategy is, when you’re first approached by a client—or you’re in the room with someone who could be a potential client—is to listen. Just ask questions. Get them to talk about the problems they experience day to day. Try to understand their operational workflows—what frustrates them, what they complain about. And then, if some of those problems are things you feel you can address efficiently—especially if you’re new to a market—just take whatever they’re willing to pay you and fix them. Once you do that, it starts building trust. And eventually word gets out, and next thing you know, you’re being introduced to other people. You don’t have to do cold calls or sales emails. You just keep solving problems and listening, and eventually you kind of work your way up the food chain.

Robert Duncan: I want to get to AI, but before I do, there are a couple of things I want to touch on from that time. Did you ever hear directly—either from Pope Francis via Cardinal Czerny, or Father Czerny at the time, or in some other way—about the work you did on the migrants and refugees side?

Matthew Sanders: I prefer not to comment. So suffice it to say, I think we did what we were asked to do, in the sense that we integrated the technologies and they became part of the day-to-day workflow. And I think at a certain point, once they were integrated, we realized that where things needed to go next didn’t necessarily have to involve us. So I felt good about what we did. And I certainly feel that, given we started working with other groups within the Vatican as well, the way the work was done and the effects of it were generally respected and well received. I did have the opportunity to meet Pope Francis at one point, but I was always very reticent to meet the pope. I generally think it’s a good rule of thumb to keep the lowest profile possible when you’re doing work here. I think that’s a good survival strategy. If you’re coming to Rome and your intention is to solve problems, I would keep that in mind.

Robert Duncan: Why is that?

Matthew Sanders: Anyone who’s worked in something like a federal bureaucracy understands there are a lot of complex interests. Ironically, in for-profit corporations—at least in my experience—it’s clear why they exist, and it’s clear what you need to do to keep your job: you have to add value and help increase the bottom line. With non-profits and government organizations, the incentives are much more complex. So when you’re talking to the head of one department or another, their incentives can be very different. There isn’t always a single unifying objective. In principle there is—let’s say evangelization—but how that’s carried out can be interpreted very differently. So you have to be careful. You might come in thinking the way to solve the Church’s problems is one particular approach, and start pushing that, but another group may have a very different approach. And you can end up alienating them or frustrating their efforts. That’s the challenge of working in larger bureaucratic organizations—navigating that complexity. Unless you have a mandate from on high, where leadership is clearly saying, “We want this done this way,” you have to be very careful.

Robert Duncan: I think what you’re describing is something you learn working in and around the Vatican—that it’s not a monolithic structure, but lots of smaller offices that are, hopefully not, but sometimes maybe working at cross purposes.

Matthew Sanders: Yeah. I think everyone believes they’re trying to serve the mission of the Church. I firmly believe that. There isn’t anyone I worked with at the Vatican who I thought was in it for themselves. But what surprised me is how differently people can interpret what that mission—evangelization—should look like on the ground. And that makes sense. You have people coming from all over the world with very different pastoral experiences and insights, and then they’re here in Rome. So they bring that past experience to bear on their current work. And that means they tend to look at things in a certain way. And without understanding where someone is coming from—what that pastoral background is—it’s very difficult to get on the same page sometimes. That’s why I think the best posture when you come here is just to listen and respond. That may mean sometimes you’re solving problems you don’t think are that important. But relationships here are paramount. So sometimes it’s better to bite your lip, do what’s asked, build the relationship, and then over time, as that relationship grows, you can begin to suggest other ideas.

Robert Duncan: So during all of this time that you’re consulting with the Holy See and institutions in Rome, I remember you jet-setting—you’d go to Davos for the World Economic Forum—and you were building relationships with executives at Google. Can you tell me about that, those background projects and personal interests?

Matthew Sanders: Sure. One of the unique experiences of being here is realizing how anyone in a position of power or influence sees the relevance of coming into contact with the Holy See in some way.

Robert Duncan: Sorry—who do you mean? People at the World Economic Forum?

Matthew Sanders: Yeah, I mean heads of global corporations, politicians, that kind of thing. Not all politicians, but certainly a good number recognize they have Catholics within their constituencies. And in the case of corporations, their brands are very important—they want to be seen as a force for good in the world. And I think most of these companies, and many countries as well, recognize that the Holy See is a very potent force for good. So it’s important for them to maintain an active dialogue with it. So that means there are a lot of very interesting people coming through Rome. And because we were working on technology at the time, there were others here working on similar questions, but from different angles. For example, Father Philip Larrey, who is the Dean of the Faculty of Philosophy at the Pontifical Lateran University, was studying the philosophical implications of AI and emerging technologies, while we were looking at how to apply those technologies in service of the Church’s mission. Cardinal Turkson eventually connected us, and we were asked to attend an event—basically to be his eyes and ears. It was an impact-focused gathering of business leaders and others here in Rome, and we were asked to listen and provide feedback. What struck me was seeing this room of very different actors all coming together, trying to figure out how their companies or countries could work with the Holy See to do good in the world. That was impressive. But what became clear to me—especially after going to the World Economic Forum—is that while almost everyone wants to make the world a better place, the real question is: what does a better world actually look like? What is a human person meant to look like?

And that’s where I felt the Church has a unique role. It has been studying the human condition for 2,000 years. In that sense, it’s the world’s leading expert on the human person. So this is an area where the Church can offer real clarity. That’s one of the reasons Father Philip and I started working together—to bring leaders from technology and government into more direct contact with representatives of the Holy See, and to have more granular conversations about all this technology we’re building. What is the telos of it? What are we actually trying to do? We need a clear sense of where humanity should be going—what “winning” looks like—so we can work backward and make sure the technologies we’re building are aimed at the right problems.

Robert Duncan: It sounds a lot like what Pope Leo is talking about now.

Matthew Sanders: Yeah, I think so.

Robert Duncan: I want to get to that, obviously, but just for the sake of detail and richness, can you give me a few examples of some of these people you worked with and what they were after specifically?

Matthew Sanders: Yeah. Over the course of a few years, we had quite a few convenings, and we were privileged to have some pretty remarkable executives come in. For example, Linda Yaccarino, who was CEO of X—formerly Twitter—came to one of our forums. Tae Yoo, who was Senior Vice President of Impact at Cisco, participated. And then Matt Brittin, who was head of Google for EMEA—Europe, Middle East, and Africa—also contributed. And that’s just naming a few. We were also honored to have the President of Malta come to one of the forums and offer insight. So it was a very rewarding experience. It made me feel optimistic, in the sense that, on an individual level, everyone wants to see humanity flourish. People really do internalize the injustices in the world. The real obstacle is that you have these different verticals—business, government, religion, NGOs—with very different incentives. And trying to align those incentives, and then figure out practically how they can work together, is very difficult. That’s not a problem we solved. But I do think some good came out of it. And even today, we continue to benefit from many of those relationships.

Robert Duncan: One thing I know about you from our past conversations is that you were thinking about AI in a serious way long before most people in the Church I knew. This was maybe seven or even ten years before ChatGPT came out. And nobody really predicted—even a year before—that we were this close to what we have today, except maybe people inside the industry. I remember long discussions with you about the moral and theological questions that would arise if AI made those advances. Can you talk about your early interest in artificial intelligence—where that came from—and, before ChatGPT, what you expected the technology to become?

Matthew Sanders: Well, I think like a lot of people, I grew up watching science fiction. And I recognized a pattern: you have science fiction demonstrating the efficacy of a certain kind of technology. If you go back to the original Star Trek, you have the tricorder and the communicator. Maybe the tricorder is still a work in progress, but the communicator—you start to see mobile flip phones emerge. And I realized, okay, Hollywood has a tremendous impact on the creative imagination of people. In some ways, it almost prophetically identifies use cases for technology. Then a few years later, people are primed—they think, “That actually looks useful. Imagine if I had that.” And then industry responds: “Maybe we should try to build this.” And before long, it’s built and in the hands of millions or billions of people. So I saw this pattern happening. Some people watch that and say, this is entertaining, and that’s it. But I looked at it and thought: if we take this seriously, and if these use cases keep showing up—especially with AI and robots, which are often depicted in apocalyptic ways—then there’s a real possibility that, out of curiosity alone, we’ll build these technologies without really thinking through whether they’re good or bad, or what their impact will be. It’s almost like Hollywood presents a problem and a solution, someone at MIT says, “That’s an interesting challenge, let me try to build it,” and then it exists. And before we’ve had time to study it, it’s already affecting us in profound ways. Because of that pattern, I assumed we would eventually have AI. But I was skeptical of the doomsday scenarios. I didn’t really understand why inventing AI would necessarily lead to humanity’s destruction.

Robert Duncan: You were pretty positive about its promise, I think.

Matthew Sanders: Yeah. And I think that’s partly because I tend to be an optimist. I think most people are trying to do good. So I didn’t see why the technology would take such a dark turn, especially since we’re building it in some sense in our own image. Obviously, it can have complex incentives, and there are bad actors who can use it in the wrong way. But the idea that the technology itself would “wake up” and decide to take us out—that just didn’t make sense to me. That wasn’t a scientific claim. It just didn’t feel like the doomsday arguments were grounded in science. They seemed more rooted in fear—fear of human beings, really.

Robert Duncan: For me, when ChatGPT dropped and I saw what it could do—we went from very basic chatbots we’d had for years, like the old AOL or Yahoo ones, to something that felt like a quantum leap. Did that surprise you?

Matthew Sanders: Yeah, it did. I’m not entirely sure why I was as surprised as I was. I think, like a lot of people, I underestimated how much of a difference scale would make.

Robert Duncan: When it first came out and it was producing these long, coherent responses, I was kind of shocked that it held together. I figured maybe it could generate sentences—we’d seen chatbots before—but the way it seemed to intuit what you were asking and almost plan its answer and then write it out… for the most part, it felt like magic.

Mathew Sanders: It really did feel like magic. And like a lot of people, I was a bit bewitched by it. I anthropomorphized it. I assumed there was more going on than there actually was. It was hard to understand how something like a next-token predictor—basically an autocomplete model—could handle language at that level and produce what looked like insight. They’re not true insights—it’s drawing from a massive humanities dataset and remixing it.

Robert Duncan: I remember one of the first things people were doing was taking Shakespeare and saying, “Do this in the style of Donald Trump,” or something. And it was remarkable—its ability to imitate. But you say there was a mystery to it, and there still is a mystery to it. Even the designers don’t entirely know how it works.

Matthew Sanders: Yeah. I think that was the thing that was most surprising. In the past, technology was very deterministic in the sense that we wrote the logic for it. So yes, it was inherently predictable, because someone wrote that code. With AI, it was very different. Anthropic has talked about this a bit—we set the conditions for its emergence, and it emerged, but we didn’t program a lot of these capabilities. They just arrived. We fed it lots of data, gave it lots of compute and an objective function, and next thing you know, it figured out how to do these things. So that was surprising—how well it worked. And it was also a bit unsettling, because this is the first time we’re building a technology where we don’t exactly know how it’s built. I mean, we do in the sense that we know how to put certain ingredients together, and if we run current through it, something emerges. But we don’t know exactly how it learns to understand concepts or perform certain skills. So in some ways, we don’t really know what the limits of this approach are, which to me is quite unprecedented.

Robert Duncan: How quickly did you think, “I need to apply this technology in the Church?”

Matthew Sanders: Pretty quickly. I had to take a look at it and say, okay, how did they do this? I need to understand this. I had been paying attention to AI, but the use cases were very narrow—classical machine learning—and it usually took a lot of effort to make a particular AI component work in a production setting. So the general use of generative AI was kind of shocking. First I wanted to understand how it was built. And then when I realized how it was built, I thought, okay, hang on a second—that makes a lot of sense. If they’re taking basically everything that’s ever been posted on the internet and training on that, and it’s producing this kind of autocomplete model, then in principle we should be able to reverse engineer that and do something similar for the Catholic Church—leveraging all the Church’s data and being able to distill it reliably. The problem was that the Church’s dataset was too small to get something like a ChatGPT 3 or 3.5.

Robert Duncan: What do I mean by dataset?

Matthew Sanders: The ChatGPT models—especially GPT-3—were trained on basically everything publicly available on the internet, the Common Crawl. It’s a vast ocean of text—everything we’ve said publicly online. That’s hundreds of billions of tokens. You need that density of data for the model to map and understand concepts, build representations, and be able to autocomplete language—understand a prompt and provide an answer. I won’t get too much into the technical details there, but—

Robert Duncan: But you do understand the technical side of it? I mean, when you and I talk, we often stay at a conceptual level about how this will impact the Church. But you run a business, so I imagine you have a technical team behind you—you understand how this works.

Matthew Sanders: Yeah, at a certain level of abstraction. I’m not down at the kernel level, optimizing kernels, for greater efficiency on AI, but yeah, I understand the ingredients in principle and I understand some of the unique challenges and what it will take to, to address those.

Robert Duncan: Right.

Matthew Sanders: It’s one of the reasons why, in addition to building systems, we also are doing kind of research as well, you know, in our own way.

Robert Duncan: So you’re talking about datasets, and there was this particular challenge that the Church had of not having enough data. That might surprise people listening, because the Church is 2,000 years old—we have a ton.

Matthew Sanders: It is surprising. It’s counterintuitive, because yes, I think we probably have more data than any single institution in the world. But it’s still comparatively a very small amount of data. If you look at the data necessary to create something like ChatGPT-3, and then you look at the Catholic dataset—especially what’s been digitized—there’s a lot more data sitting on shelves that hasn’t been extracted and can’t yet be used. So what we have available is woefully insufficient to get those emerging capabilities. That’s not to say you couldn’t train some kind of basic AI system or a basic autocomplete model, but its generalizing capabilities would be extremely limited. Its ability to intuitively understand your prompt would suffer because of the size of the dataset. A lot of the reasoning capability and intelligence of the model comes from having studied a vast amount of text and how it all relates. So the smaller the dataset, the less conceptual understanding it has of language and, to some extent, of the world. It would be like the difference between interacting with, say, a PhD student—which is roughly where modern AI is now—and a model trained only on a small dataset, which would be more like talking to a four-year-old who has learned the basics of the Catholic faith.

Robert Duncan: An advanced four-year-old!

Matthew Sanders: Maybe a slightly advanced four-year-old, but still. It’s not that it doesn’t have access to the knowledge—it’s that its ability to properly understand what you’re asking and give you the answer you’re actually looking for requires a certain level of intelligence. You and I have grown up and developed a complex understanding of the world, and because of that, we can assume a lot about what’s being said. Language is an abstraction of reality. So if you train a model on a very small amount of language, it ends up with a very abstract and limited understanding of reality—and that makes it frustrating to work with.

Robert Duncan: So how’d you fix this, or how are you fixing this?

Matthew Sanders: Well, initially it became very clear to us that eventually the Catholic Church needs to train its own AI. So we said we’re going to have to have our own training program—we can talk more about that, which we do. But we realized we had to find some way—really the question was: if we can’t train our own ChatGPT-3, can we find a way to make it safe? Because right now, two of the major challenges with ChatGPT-3—well, there are more—but one is its tendency to hallucinate, to make things up.

Robert Duncan: They’ve gotten better about that, right?

Matthew Sanders: They have gotten better, but hallucination is still a big problem. And by nature of their training, the models don’t always know why they’re saying what they’re saying, or where it’s coming from.

Robert Duncan: Do they ever know why — well, that’s different—that’s another question.

Matthew Sanders: Right. But what I mean is, when a model is generating, it doesn’t always know what it should be citing, at least in terms of priority. And even citing—saying, “Here’s the source document I referenced to arrive at this insight”—it can do that to some extent, but it’s not very reliable. So there’s this tendency to make things up, because it’s eager to please, and then there’s the issue of lack of transparency. You don’t really know where its outputs are coming from. Is that insight coming from a Reddit thread, or from an encyclical? The model itself doesn’t really know. So we needed to figure out how to leverage its mastery of syntax—its mastery of language—but ensure that everything it’s saying is grounded in primary source documents. And we had to find a way to minimize hallucinations. That’s where the research project, Magisterium AI, came from. We wanted to see how far we could go using scaffolding—different approaches that essentially fence in the model, keep it on rails. It took us about six months to figure out the right initial configuration. Then we needed more people to test the system. That’s when we did an interview with EWTN, and it kind of went viral, which was a big shock to us. We couldn’t get back online—there was too much traffic. And Father Philip had to reach out to Sam Altman to help us increase our rate limits. It was a pretty crazy time.

Robert Duncan: So you had Sam Altman—for people who may not know, he’s the CEO of OpenAI, which created ChatGPT. So you were working with the top?

Matthew Sanders: Well, I should say, he was gracious enough to understand the importance of what we were trying to do. And he knew, given how small we were, that we wouldn’t be able to be prioritized from a compute perspective. So he was willing to step in and get us the capacity we needed to continue the project, which I’m very grateful to him for. It’s rather ironic, but if it wasn’t for Sam, we wouldn’t have been able to get back online and continue the work. And really, the work at the time was just trying to get as many philosophers, theologians, and canon lawyers as possible to test the system, to expose its vulnerabilities and deficiencies, and figure out what more we could do to make it more reliable. That’s eventually what led us to digitizing everything the Church has said, to grow the Catholic dataset and provide access to more. But that’s a more recent project.

Robert Duncan: Right. So you launch Magisterium AI. Just tell people what that is. You actually went online, and it was called Magisterium AI.

Matthew Sanders: Yeah. What we ended up building, we call a compound AI system. For most people, you go to ChatGPT, and you interact with ChatGPT through the user interface. In the same way, we built a user interface for people to interact with our compound AI system. So it operates just like ChatGPT. You go and ask a question. The difference between Magisterium AI and interacting with ChatGPT directly is that if you ask ChatGPT a question, that question goes to the model. The model does a lot of highly complex geometry and math, and then it basically does next-token prediction and generates an answer. You’re going directly into the model’s brain, and it’s then spitting out an answer.

Robert Duncan: And for people who don’t know—Magisterium—what does that word mean?

Matthew Sanders: Oh, Magisterium?

Robert Duncan: Yeah.

Matthew Sanders: So Magisterium is the teaching authority of the Church. So it’s basically everything the Church has formally taught—that’s all encompassed within the Magisterium. And of course, the bishops and the Pope are the primary communicators of the Magisterium.

Robert Duncan: So for people coming into this cold, that means all the dogma and doctrine of the Catholic Church—official teaching—your AI system is meant to plumb those depths.

Matthew Sanders: That’s right. So essentially the project was: could we take everything the Church has said formally—all the doctrines and dogmas within the Magisterium—as well as notable scholarly work from the doctors and fathers of the Church, like Thomas Aquinas and St. Augustine. They’re not all formally part of the Magisterium, but they’re still very important voices. So we wanted to include their insights as well. The idea was to digitize all of this data and make it accessible to anyone in the world, on any device, in their native language. For me, becoming Catholic was a real challenge. It was difficult to understand the faith—I didn’t understand what a Franciscan was, I struggled with certain doctrines. It took a long time to piece it together. So I thought, I wish there were some kind of resource where any question I had, I could find an answer quickly and efficiently. That’s what Magisterium AI is meant to be. You ask a question, it looks across all these source documents—everything the Church has said—finds what’s relevant, and then distills that information so it’s accessible.

Robert Duncan: How many people are using it?

Matthew Sanders: Millions at this point.

Robert Duncan: Millions?

Matthew Sanders: Yeah. So there’s Magisterium AI, which is our own ChatGPT-like product—our user interface. The mission of the company is building and scaling Catholic AI. We spent a lot of time on the building part, but then the question became: how do we scale it? Because building Catholic AI is complicated—it takes a lot of energy and resources to do well. And if every organization had to climb that same mountain, scaling would take forever. So we decided we needed to let people build on top of us. That’s why we created an API—an application programming interface—which allows other organizations to use what we’ve built. For example, Hallow, the world’s number one prayer app—they use our API. We work under the hood in their chat application. So when someone asks a question in Hallow, they can be confident the answer is grounded in primary sources, provided by us. So by building Catholic AI, making it available through Magisterium AI, and allowing others to build on top of it, it enables the Church to leverage faithful AI faster.

Robert Duncan: Do you have age and demographic data on the people who are using it?

Matthew Sanders: No, not really. People can voluntarily provide that information if they want. We do have a personalization component to Magisterium AI. So if you provide that information, it helps, because the AI has more context. If you’re 13 years old, for example, it’s helpful for the system to know that, because it can try to explain things at that level. But it’s not required. Users are anonymized, and that’s intentional, because a lot of people are asking very personal questions. It’s safer for us to keep users anonymous out of respect for them. We do keep data to improve the product, but we don’t share that information without explicit permission.

Robert Duncan: So you can see what people are asking, but you don’t know who. What are some of the trends and patterns? The Church, especially in the United States, is interested in reaching people who have disaffiliated—the “nones.” As people interact with Magisterium AI, you must be getting real insight into their hangups, their questions, their points of confusion. What have you learned?

Matthew Sanders: The short answer is, I’ve learned a lot. I had some instincts about what people would ask, and many of those were confirmed. You see questions ranging from very big ones—what’s the point of the universe—to very personal struggles, like someone dealing with a particular issue, maybe even something like a sexual addiction. People are asking, they’re struggling, they feel lost or discouraged, and they’re looking for insight—from the popes, from the fathers—to help guide them. Those are all things we expected. But what surprised me is that when we first launched, we were focused on building a tool for pastors, for people working in chanceries, for catechists. So I expected more straightforward doctrinal questions—“What does the Church teach on this? Give me the sources.” And that’s still one of the most common types of queries. But what surprised me was how quickly it was adopted by the laity, and how many of the questions were highly pastoral in nature—which was a challenge for us when we first started building the product.

Robert Duncan: Like, “I have a friend who”—or “I have a family member”—”and I want to reach them”?

Matthew Sanders: Yes. Or, “I’m struggling with this particular thing,” or, “Here’s something that happened in my life and I’m not sure what the right thing to do is.”

Robert Duncan: Why do you think people are going to a machine to ask those questions?

Matthew Sanders: My intuition is it’s because they’re not finding someone in their immediate orbit who has studied the faith well enough to answer those questions. For most people, those people just aren’t there. Priests are extremely busy. People know they could ask their priest, but it’s difficult to get ahold of them, or they feel like they’re going to be judged. And so talking to some nameless, faceless machine feels safer than talking to a human.

Robert Duncan: I think that’s part of the core objection you hear. On one hand, these tools are solving a real problem—people are alone and don’t have anyone to talk to about these things. On the other hand, they may encourage that isolation, because some people who might have gone to another person now take the easier path that doesn’t require vulnerability. How do you see that?

Matthew Sanders: It is a problem. It’s a profound problem. Ideally, I don’t think Magisterium AI should have to exist. I wish it didn’t need to exist. I made it, yes—but mainly because there’s a bigger problem than that one, which is that people have very little hope of reforming their lives or ordering their lives if the truth is not available to them. That was my primary concern, because I was in that place. I was trying to make sense of the world, trying to figure out my purpose, and it was very difficult to get good answers. Even when you met someone who purported to represent the Church, what they said wasn’t always accurate—and I had no way to verify it. It was really hard. So what’s most important is that the truth is readily available to everyone. That’s one of the reasons Paul and the apostles went out. They didn’t spend a lot of time with every community—they couldn’t. They just wanted people to have the kernel of the truth, to point them in the right direction, and trusted that God would work through that. That’s how I see Magisterium AI. We’re trying to digitize everything the Church has said and make it accessible. Now, this other problem—people feeling isolated, fearing judgment, not trusting others enough to open up, not being well catechized—that’s real. And if those problems didn’t exist, there would be much less need for something like this. It would just be a research tool. It wouldn’t be something people sit and talk to about their problems all day. So it’s important to say: we didn’t build this to replace pastors, or catechists, or parents, or friends. We built it as a tool—to distill the wisdom and insight of the Church and make it accessible. And the hope is that by making it available, more people will take it upon themselves to learn the faith, internalize it, and live it—and that will make them more effective at evangelization.

Robert Duncan: Do you have any way of tracking whether the people who come and ask these questions—and get answers from Magisterium AI—actually advance? Whether their problems are resolved? You know, some people can be abusive to AIs, saying, “you’re stupid,” or swearing at it. I’m not sure how often they say, “thank you, you’ve answered my question, I’ll return to the Church now.” But do you have any way of tracking how effective it is?

Matthew Sanders: We do. I don’t know if I did a better job of explaining this before, but the users are anonymized. That said, we pay very close attention to the exchanges, because we’re always trying to evaluate the AI’s answers. We’re asking: could we have improved that answer? Could we have used different sources? Are there gaps in the knowledge base we need to fill? Did it have the right tone? Did it understand, in that moment, that the right course of action was to recommend speaking to a priest, rather than giving a direct answer? Sometimes it’s better not to give an answer, and to let a priest handle it, because there may be pastoral sensitivity around the issue. So we’re always trying to optimize the system. And for that reason, yes—we study the exchanges, and we see patterns. We see people come in basically yelling at Magisterium AI. They heard something on a podcast—“the Church says this”—they Google Catholic AI to check it, they find Magisterium AI, and now they’re unloading: “how wrong you are, is this true?” What’s fascinating is watching how, over the course of weeks, those conversations change. They move from that initial anger into something more fruitful. They start digging in. The conversation shifts from very specific objections to curiosity. And then eventually, you start seeing questions like: what does it mean to be a Catholic? How do you become a Catholic?

Robert Duncan: Really?

Matthew Sanders: Yeah. And then we also ask users to fill out surveys at certain intervals—after, say, 50 exchanges: “how’s your experience so far?” Then again later: “how’s it going?” We track those metrics over time. And at those points—and sometimes just spontaneously—we get feedback like: “thank you, I’ve entered the RCIA process, and Magisterium AI was very helpful.”

Robert Duncan: That’s the process to become Catholic?

Matthew Sanders: Yes. We get a lot of that. A lot. We also get feedback from parents, from catechists, from teachers preparing lesson plans. That’s what motivates me. This is a difficult project. You have people on the left who think it’s too conservative, people on the right who think it’s too liberal. Addressing that can be spiritually draining. But when you look at what users are actually saying—what their experience is—that’s what keeps me going. And also their willingness to provide feedback, to help improve the system. They want it to be better. That’s very encouraging.

Robert Duncan: So that leads into what you’re working on now. You just had a conference last fall called the Builders Forum. It was multifaceted, but one of the questions it tried to answer was how to off-ramp people—so that when they get to the point of wanting to enter the Church, you can guide them on how to do that and where to do that. So that’s a more real-world application. You think of AI as a technological, virtual service, but you’re adding to that the service of how to put people back into the world.

Matthew Sanders: Yeah. The off-ramping problem is one of the problems that preoccupies our time the most. As I said before, Magisterium AI was built to help distill the truth for people—to get the answers they’re looking for. And it’s important to note that when you’re speaking to it, you’re not really speaking to the AI system. Who you’re dialoguing with is the popes, the doctors and fathers of the Church. The AI is not rewarded for novel insights. We don’t want it accessing its neural net and providing its own ideas. It operates under very specific constraints. When someone asks a question, one system retrieves relevant documents, and another reads and distills them. The insights come from human beings. But the off-ramping problem is about recognizing when someone has, in a sense, intellectually converted. At that point, they might ask: what’s next? I’m interested in becoming Catholic—what does that mean? The AI can explain the OCIA process. Then they ask: how do I do that? You need to find a parish. How do I find a parish? Some people don’t even know to ask those questions. So we want the AI to be able to suggest it—to say, have you thought about speaking to a priest? Then the challenge is: how do we move them from Magisterium AI into a real parish? And is that always the right step? Sometimes it might be better to connect them to a lay movement or a community where they can engage with real people. That’s what we’re trying to figure out. The Church has been working on this for a long time, so we’re identifying partners—places and people we can connect them to. Then ideally we stay in touch and see how it goes. Based on that, we can determine whether we need to build additional tools, or whether there’s pastoral intelligence we can pass back to bishops and priests to help them be more effective. That’s one of the next phases—taking the insight we’ve gathered and making sure it’s fed back to the shepherds.

Robert Duncan: So maybe we should close by talking about Pope Leo. When he chose his name, he mentioned the challenge the Church is facing with artificial intelligence. What did his election—and that choice of name—mean to you personally?

Matthew Sanders: I was greatly relieved. Mainly because I have a fairly strong conviction about where this technology is going and the kind of impact it’s going to have on society. It’s going to be profound. And I worry that if this transition into a new age isn’t managed properly, there could be a lot of fallout.

Robert Duncan: And everybody’s talking about all this—if you go on YouTube or Spotify, all the podcasts are warning. It sounds like a lot of the warning is coming from the people who are actually creating AI.

Matthew Sanders: That’s right. And this is what’s very confusing. The people who are closest to the technology, for the most part, are setting off flares. They’re saying: listen, we see where this is going in the next six months to a year, and we don’t see a world where this doesn’t start causing significant unemployment—displacing a lot of human beings. And I think a lot of us still think of AI as this disembodied reality, but now we’re starting to embody it in robots. So this isn’t just a white-collar issue. It’s a blue-collar issue too. This impacts all labor in one way or another. Then you have another group saying: no, it’s going to be fine. This is just like the Industrial Revolution. Yes, there will be unemployment in certain sectors, but new jobs will be created—it’ll all balance out. Now, over certain time horizons, maybe that’s true. There was a moment where hiring for developers dropped close to zero—companies just weren’t hiring anymore. And now hiring is going up again. That’s confusing. If AI is replacing people, why are companies hiring developers? It’s because AI is starting to eat into other fields, so companies need developers to build the AI and the scaffolding around it. So yes, it is displacing work—it’s just not displacing developers right now, because we need them to build the systems. It’s confusing because you have these two narratives. You have people—especially in corporations—who want to underhype it, because they don’t want to rattle investors. You have people in government who don’t want a strong public reaction, because they don’t yet have answers, so they kick the can. What concerns me is that we should be listening to the people closest to the technology. And I think in the next two, five, and ten years, there will be waves of significant economic realignment. I think the Church, like in the past, is poised to help shepherd humanity through these transitions. But her intellectual and spiritual formation is going to be needed more than ever. So to have a Pope say, “I recognize the world is about to change, and I’m committing my papacy to ensuring the Church is pastorally ready to respond”—that was very encouraging.

Robert Duncan: Do you think the leadership of the Church—from universities all the way through the clergy—faces a particular challenge in understanding this as a major issue? Because they might assume, rightly or wrongly, that their livelihoods won’t be affected. Priests aren’t going to be replaced by robots. We’ll still need theologians, people working in Caritas, service roles. Is there a mentality that this issue doesn’t really touch them as much?

Matthew Sanders: I mean, I think from a jobs perspective, you’re absolutely right. I don’t think any of the pastoral workers per se are really at risk of being unemployed by this. I think there are certainly people working in administration in the Church who have to recognize that if the Church adopts this technology—which would make sense, because it would free up more time for pastors to focus on pastoral work as opposed to administration—there is some risk of disruption there. But I generally agree. I think more than anything, it’s just—they’re not really aware. Maybe “ignorance” is too strong a word, but they’re just not aware. And that’s largely because of how they were educated. They’re philosophers, theologians, pastoral workers. I think it’s very easy for them to look at this technology and say, “yeah, I’ve been around 70 years, I’ve seen lots of these technologies—this is just another one.” And they don’t understand how different this is, how extensive it’s going to be. And if they think they’re busy now—I think in 10 years they’re going to be a lot busier. Because there will be a lot more people having existential crises. People who no longer have a job to go to. Maybe the government provides basic needs—they have a roof, they’re not starving—but they don’t know what to do with themselves. The things they spent their whole life cultivating—the gifts they thought they had—are no longer needed by the economy. So what do they do with their time? And I think a lot of people, because they feel ill-equipped to answer that question, will turn to distraction. They don’t have good answers, so they just keep themselves distracted. And I think that’s the cataclysm we’re potentially heading toward. If we can’t find a way, at scale, to provide human intellectual and spiritual formation—to help people understand who they are, remind them of their dignity, help them discern their gifts, their vocation, and find ways to contribute to their communities—then we’re going to be dealing with a lot of people going off an existential cliff.

“If we can’t find a way, at scale, to provide human intellectual and spiritual formation … a lot of people going off an existential cliff.”

Robert Duncan: Pope Leo chose his name in part because of the AI challenge. You’ve developed an AI system that reaches millions of people. Has he reached out to you? Have you reached out to him?

Matthew Sanders: No—he hasn’t reached out, and I haven’t reached out. I think he’s doing exactly what he needs to be doing. I like to think that what we’re doing is directly responding to his call. He’s cautioned about the use of AI—we have to be very prudent. We can’t anthropomorphize it. We can’t rely on it as a crutch. For instance, priests are very busy, and there may be a temptation to use something like Magisterium AI to generate a homily on the fly, hit print, and read it. But that’s not what it was created for. It’s meant to augment pastoral workers. A good homily, if it’s really going to land, has to reflect the unique pastoral reality. Magisterium AI can help by drawing on what the popes, the doctors, and the fathers have said—that’s valuable. It allows priests to access that easily and weave it into their homilies. But it’s still a high level of abstraction. If a homily is going to land, the priest has to take time to shape it. So I think he recognizes that these tools can support the Church’s ministry, but we have to be cautious. We need proper formation and training so people know how to use them well. I like to think what we’re doing is in service to the Church. We pay close attention to what he’s signaling and try to build in a way that he would support. We also work with people in the Church—we have a theological advisory committee to ensure we’re focused on the right problems and improving the product in the right ways. But he’s a busy man. I don’t think he has any immediate need to talk to me.

Robert Duncan: Besides the words of caution that he’s outlined, what else specifically have we gotten from him so far? What indications has he given for where he might take the Church on this issue?

Matthew Sanders: Well, I think more than anything else, he’s mentioned before—and I’m paraphrasing—but he believes that building this technology in the right way is like participating in a divine act of creation.

“(Pope Leo) believes that building this technology in the right way is like participating in a divine act of creation.”

It’s almost a sacred responsibility to use the tools available to us to serve the mission of the Church, to help the Church in its mission of evangelization. So I think what you’ll see is that the real area we need to focus on right now is helping—or reminding—humanity what life is supposed to be about. Because the way we understand life is about to be disrupted. And I think in a good way. For too long, the economy has looked at us as widgets—part of a larger machine, contributing to that machine running, whose purpose is to generate profit or economic output. But that’s an unhealthy anthropology. And I think what he can do, in his role—and given how respected he is, and how many people are listening right now because they sense the disruption coming—is to re-baseline what the human condition is supposed to be. To remind people of their dignity. To remind them that just because AI and robots are doing more of these GDP-type jobs, that doesn’t mean there won’t be work to do—work we can choose to commit ourselves to. And that we can refocus on what gives life meaning: growing in our relationship with God, serving others, cultivating our vocation. So I think he’ll remind us that this could be one of the greatest opportunities we’ve ever had to rediscover our humanity. I hope he focuses on that, because I think that’s what’s needed. Even large corporations and AI labs are looking to him for that kind of leadership. And then, as we become more aware of how these tools can add value, I think he—and the dicasteries—can point to specific technologies and say: we think these can be helpful in this way, and we encourage you to use them.

Robert Duncan: Matthew Sanders, thanks for sitting down with Catholic News Service.

Matthew Sanders: Thank you.