5 Things Catholic Parish Leaders Can't Miss in the New AI Encyclical

Your new AI encyclical and your church software are about to have an argument on Monday morning. Most parishes will referee it wrong.

Picture the staff meeting. The pastor sets the document on the table and asks the room, "So what are we supposed to do about this?" Then he looks at the one person who actually runs the church-management system. That person is already behind. Duplicate profiles in the database. Year-end giving statements due. A new-visitor list nobody has called.

One reading says put the tools down and do it by hand, like you mean it. The other says the Pope is asking something harder than "use it" or "don't."

Magnifica Humanitas is not a software ban, and it isn't a buyer's guide. Pope Leo XIV writes that "in the era of artificial intelligence, when human dignity is threatened by new forms of dehumanization, ours is the pressing duty to remain profoundly human" (#15). That's not a rule about which app you may install. It's a duty about who stays in charge while you use it.

Nine Catholic AI experts spent three long conversations working out what that means on the ground, three panels from the Faith & AI Project: Truth, Work, and Freedom, the Culture of Power, and Technology and Dominance. They followed a same-day livestream reacting to the document the morning it dropped. Five things kept surfacing. Every one comes down to the same muscle: judgment. Knowing when to take what the software hands you, and when to overrule it.

1. You can't automate a conscience

Your software can sort your parishioners. It cannot be responsible for them.

The encyclical names the danger exactly: handing a system "the power to select who is worthy or not, without anyone bearing responsibility for that judgment" (#103). Your giving dashboard can flag a lapsed donor. Your engagement tracker can rank who's drifting. Both are useful. But the software never decided that a family wasn't worth a call. You did, or you didn't. The recommendation is the machine's. The answer for that family is yours.

Fr. Philip Larrey, who taught philosophy in Rome for years, put it as plainly as anyone has:

"In order to make a moral decision, you need a conscience. You need responsibility, you need a life, and you need accountability. AIs don't have that."

–Fr. Philip Larrey

2. The tool already has a point of view

There is no neutral church software. Every tool ships with somebody's priorities baked in.

"Technology is never neutral," the Pope writes, "because it takes on the characteristics of those who devise, finance, regulate and use it" (#9). You see it in the small defaults. The auto-suggested "we missed you!" tone that reads as glib to a family that just buried someone. The visitor-note summary that keeps the email address and drops the fact that she asked about the annulment process. The question isn't "when do I put the tool down." It's whose choices and priorities are inside this thing, and do they line up with mine? Override the defaults that don't. Make the templates sound like your parish, not a generic SaaS welcome flow.

Taylor Black, who works on AI at Microsoft and runs a Catholic AI institute, says it from inside the industry:

"Every time we ship a generative AI product, we're implicitly shipping a certain understanding of human anthropology, and tech is not the authority on human anthropology."

–Taylor Black

3. A score is the start of a conversation, not a verdict

An engagement number tells you where to look. It does not tell you what's true.

"The quality of a civilization," #114 says, is measured by "its ability to recognize the other as a face not merely as a function." Here's the override in practice. The system marks a parishioner "low engagement, low priority." But a staff note from three weeks ago says his wife just died. The data is right and useless at the same time. You call him anyway, and you call him first. The score points; the person decides; the person wins the tie.

Brett Robinson, who studies this at Notre Dame, warned that we've collapsed dignity into something thin:

"We've fixated on something like social dignity. That's a very thin and shallow form of dignity. It's not the ontological dignity given to us by God."

–Brett Robinson

4. Name what you're handing off

Decide, on purpose, which work is the task and which work is the relationship.

The encyclical says "every technology shapes those who use it," and that real formation means learning "when and for what purpose it ought not to be used" (#140). Run your office through that test. The standing tasks first. Scheduling, reminders, receipts, duplicate cleanup. Hand them over without guilt. Then the other column. The first call after a death, sacramental prep, walking a returning Catholic back through the door. That's the relationship wearing a task's clothing, and it stays yours. Sort the two now, on a calm afternoon, before a busy Tuesday sorts them for you.

Fr. Jean Gové, who coordinates AI research for the Vatican's culture office, gave two questions that work as a field test:

"What am I offloading onto the machine that I should be doing? What am I getting from a tool that I should be getting from a relationship?"

–Fr. Jean Gové

5. Slow is a feature

Just because the system can send it instantly doesn't mean it should.

"Let us cultivate relationships!" Leo writes. "In an era that favors speed and fragmentation, the human person still yearns to receive care and recognition from attentive minds, kind words and hands capable of tenderness" (#239). Automation sells speed, and most days that's a gift. But a grief follow-up that auto-fires the same afternoon as the funeral is the wrong kind of fast. Personalized, instant, efficient, and exactly wrong. Some sequences should carry a human pause on purpose. Knowing which ones is the whole discipline.

Julianne Stanz, who works in evangelization for Loyola Press, named the discipline:

"This document is calling us to a particular kind of pace, a slower pace, a discerning pace, a pace that listens to the Holy Spirit."

–Julianne Stanz

So which is it?

Notice what the five have in common. None of them is a case for ripping the data out. Read together, they're a single test: are your tools keeping a human in charge of the judgment, or quietly making it for you?

That test cuts both ways. If a human stays on the judgment, the software stops being a threat and starts being a help. The dashboard that flags a drifting family isn't replacing your pastoral care. It's handing you a fact you didn't have, so your care lands where it's needed. The danger was never the data. It was letting the data decide.

So, back to that staff meeting. Rome isn't asking you to do your data entry by hand to prove your devotion. It isn't blessing whatever the software defaults to, either. The encyclical settles the argument by changing the question. Not "use the tool or don't," but "who's holding the judgment?" Keep the tools. Stay the shepherd in the office.

And that answers the harder one. Treat the "who to follow up with" list as a draft and never a verdict. Read it, add what the software can't see, decide for yourself who gets the call. Do that, and the dashboard stops burying the face. It starts finding the one you'd have missed: the name that slipped, the family that drifted, the person who needs a call this week. The software clears the fog. You still have to look.