AI helps me write this newsletter. To press mute on the alarm bell that may have just gone off somewhere, here’s my position: it’s an instrument, not a co-author. I stand behind everything published here, from the judgment and the theology to the decision to say this and not that.
While how much I lean on AI is an ongoing discernment (and I won’t pretend otherwise), there’s a hard line I won’t cross on principle: nothing here substitutes for the parts of ministry that can only happen in person—that conversation after Mass, a visit to a hospital room, a confession. This newsletter can carry what I preach a little further than the pulpit reaches, but it can’t replace my actual voice in those moments.
So why use AI at all, in a newsletter called A Middle Manager’s Guide to Parish Ministry? Because it’s a very middle-manager thing to do.
How an Issue Gets Made
Middle managers aren’t the ones wordsmithing the company’s values, mission, and vision statements. They receive those documents from above, internalize them, and translate them into something the people in the field can act on, with an eye toward efficiency the whole way through. A pastor works in the same spot. The faith reaches him already settled; his part is to teach it and put it into practice in a parish. The first issue called that order Creed before Mission. For this newsletter, AI helps with some of that process. Not all of it.
The receiving is where AI does a lot of visible work. It helps me research and check anything that touches doctrine or fact. Every claim about Church teaching gets run past a Catholic research tool that answers from the Church’s own documents and cites them, along with a general check against primary sources. Scripture is checked word for word against the New American Bible, Revised Edition, and quotations against the text they come from.
Dates get checked for when as well as whether: an early draft of the first issue had the Creed prayed at Mass ever since Nicaea, when Rome didn’t adopt it until 1014. The tool gets things wrong too, which is why it never gets the last word. That belongs to me.
The internalizing is where AI has no part at all, and I wouldn’t want it to. Receiving the Church’s teaching is one thing; letting it take root in me is another. That happens slowly, in prayer, at the altar, and in the kind of discernment that can’t be hurried or handed off. Luke says Mary “kept all these things, reflecting on them in her heart” (Lk 2:19). No tool keeps anything in a heart. That part is mine alone.
The translating happens twice. The first translation is parish ministry itself, long before anything reaches a keyboard. Explaining one of his parables, Jesus puts it plainly: “The field is the world” (Mt 13:38). For a pastor, the field is a parish: a finance council meeting, a family planning a funeral, a new staff member asking how things are actually done. That’s where doctrine, Scripture, and the best of management wisdom become ordinary parish life and get tested against it. The second translation turns those experiences into this newsletter’s field notes, and that’s where AI helps the most.
And then there’s efficiency. Middle managers optimize, or at least the good ones do, in my humble opinion. Optimization isn’t required in parish ministry (actual parish ministry is messy and disorganized more often than not), but it is required here. This newsletter isn’t my job, or even my second one; it’s something I do at the edge of a full parish workload.
So here’s a second line, as firm as the one I drew at the start: I won’t allow this newsletter to starve my actual ministry. That’s also why AI handles most of what surrounds the writing. It made the website’s artwork, wrote most of its code, and helps me fix what breaks. Every hour it saves there is an hour the newsletter doesn’t eat from my parish. If your parish office is weighing the same tools, those two lines (what AI never replaces, and what it’s never allowed to crowd out) are a fair place to start.
In the writing itself, the tool saves time at both ends: less digging through a library on the receiving end, and less time turning what I’ve already lived into paragraphs on the writing end. It doesn’t generate what I think, and the phrasing it helps with won’t be original, which is fine, since originality isn’t in a middle manager’s job description, and it isn’t in AI’s either.
There’s real scholarship on what happens to a body of writing when everyone leans on the same tool the same way. In a 2024 study published in Science Advances, Anil Doshi and Oliver Hauser had people write short stories, some with AI-generated ideas to draw on and some without. The AI-assisted stories were rated better written and more enjoyable, with the biggest gains among the less creative writers, but they also came out more similar to one another.
AI, when left to its own devices, drifts toward the statistical middle: safe vocabulary, predictable rhythm, sentences that could have come from anyone. Quite fitting for this newsletter: mid-level writing from a mid-level manager. But middle-of-the-road phrasing isn’t the same as middle-of-the-road substance. The ideas here went through the first translation, in homilies, meetings, and hard pastoral conversations, before they ever showed up as paragraphs.
The Biblical Tradition
Handing off part of that second translation, the mechanics of getting it onto the page, isn’t new. Paul did it regularly. Romans ends with a greeting from a man named Tertius, who identifies himself as “the writer of this letter” (Rom 16:22). Elsewhere, Paul draws attention to the moment he picks up the pen himself, as though it’s the exception rather than the rule: “See with what large letters I am writing to you in my own hand!” (Gal 6:11); “This greeting is in my own hand, Paul’s. This is the sign in every letter; this is how I write” (2 Thess 3:17), and again in 1 Corinthians and Colossians (1 Cor 16:21; Col 4:18).
His usual practice, on the evidence of the letters themselves, was to dictate the substance and personally sign off at the end. I’m not claiming any equivalence here—nothing I write for this newsletter is inspired Scripture, and I’d ask you not to hear it that way. What I am claiming is narrower: whose hand held the pen doesn’t disqualify who the letter is from. That distinction is as ancient as Scripture. However, AI does more than take dictation, which is why the next example matters more.
An entire Gospel rests on the same principle, and its writer did more than take dictation. Mark wasn’t one of the Twelve and, by the earliest account we have, never followed Jesus during his ministry. That account comes from Papias, a bishop writing early in the second century, who says he had it from an earlier teacher he calls the Elder; it survives because Eusebius preserved it.
According to Papias, Mark became Peter’s interpreter and carefully wrote down what he remembered of the Lord’s words and deeds, though not in order. Peter had taught as his hearers needed, without trying to give a connected account, and Mark took care to leave out nothing he had heard and to add nothing false.
Irenaeus, a generation or two later, likewise ties Mark’s Gospel to Peter’s preaching. The Church has never treated any of this as an authorship problem. Mark’s name is on the Gospel because Mark did the writing, and Peter’s preaching is underneath it because that’s where so much of the substance came from. Neither fact embarrasses the other.
And underneath all of it sits the plainest fact: Jesus never wrote a Gospel. The only writing of his that Scripture records was on the ground in the temple area, and the Gospel never tells us what he wrote (Jn 8:6, 8). Everything else we have of him passed through other hands first: memory, preaching, translation, four evangelists, each selecting and arranging under the Spirit’s inspiration. Far from a defect in the Gospels, it’s how a message that important made it from a hillside in Galilee to you, reading this on a screen.
Mediated transmission isn’t a crack in the Christian tradition. It’s what the word means: tradition comes from the Latin traditio, a handing on.
Nove, non nova
This tradition of transmission finds an apt image back in Matthew’s thirteenth chapter, the same one where the field is the world. Jesus closes that run of parables with a mini-parable only a sentence long: “Every scribe who has been instructed in the kingdom of heaven is like the head of a household who brings from his storeroom both the new and the old” (Mt 13:52).
For most of my priesthood, the old has been the faith brought out in person—from the ambo, at a hospital bedside, across the desk in the parish office. This newsletter is the new, the same faith from the same storeroom, brought out in writing for readers I would never otherwise reach. AI belongs to the bringing-out, as one of the means of carrying things from the storeroom, never as one of the things carried.
Vincent of Lérins gave the rule for this in the fifth century: teach what you have received, and when you say it in a new way, don’t say new things (Commonitory, 22). His phrase survives as a shorthand: nove, non nova—in a new way, not new things.
That brings us back to “originality.” A scribe instructed in the kingdom is praised not for manufacturing something new from nothing but for stewarding what’s already been given, knowing how to bring it out at the right moment, and arranging it because he’s the one who actually knows the household. That’s a fair description of what priests have been doing in their preaching for so long that the practice is almost taken for granted.
A preacher worth his collar is always a little bit of a thief, borrowing a line here, an insight there, from every book and homily that ever moved him, and handing it on as if it were his own, because by the time it reaches the ambo, in some sense, it is. I’ve done that for years without a second thought, and I don’t hold this newsletter to a stricter standard of originality.
But I do hold it to a stricter standard of credit. A homily is spoken to one congregation on one Sunday; a newsletter is published, forwarded, and archived. So here, a borrowed line gets its quotation marks and its source, and the notices at the bottom of each issue say whose words I’ve borrowed.
Distinct, Not Blended
On the About page, I mention a habit I’ve half-jokingly named the Chalcedon dynamic: reaching for the shape of a doctrine and asking what it has to say about whatever is sitting on my desk. This week, what’s on my desk is the newsletter itself. Chalcedon confessed Christ as one Person in two natures, without confusion, without change, without division, without separation. I’m borrowing that shape, not transplanting the doctrine, and the analogy ends where Christ’s uniqueness begins. Here, the two are my contribution and the tool’s, and what holds them together is a single voice: mine, writing as Titus Bellwether.
Applied to how this newsletter gets made, that doesn’t mean every paragraph is some negotiated percentage of human and AI content, a 60/40 split I’d have to disclose like a nutrition label. It means, on the one hand, that AI supplies real, instrumental service—drafting, structuring, helping me get material out of my head faster than I’d manage alone—without being confused with a co-author with its own convictions or its own stake in what’s true.
And on the other hand, at the same time, it means that I’m the one who supplies the judgment, the verification, the theology, the responsibility for every word that goes out under this name, undiluted by the process that got it onto the page. Not blended. Distinct, and unequal: the tool serves; I answer.
So this is my confession, in the sense the title borrows from books like David Ogilvy’s Confessions of an Advertising Man: not an apology, but an insider’s plain account of how the work actually gets done, offered because it’s true and worth saying. The new and the old, one scribe, one storeroom. I think that’s a more honest description of this newsletter than “100% human” or “100% AI” could manage on its own.