Sermon Prep
If AI Writes the Sermon, Is It Still Mine?
Say “AI wrote my sermon” out loud to most pastors and you’ll get a wince, not a shrug. That reaction isn’t laziness-avoidance or tech-anxiety. It’s identity. A sermon is supposed to be what happens when a pastor wrestles with a text on their knees — not a finished product handed over by something else. That instinct is correct. And one of evangelicalism’s sharper voices has already put it into words.
The objection is real — and it’s right
Trevin Wax, senior editor at The Gospel Coalition, made the case directly in an April 2025 piece titled “The Missing Heart in AI-Generated Sermons.” His argument: AI cannot worship. If the point of preaching is to behold the glory of Christ and stand in awe before his grace, then a sermon produced by something incapable of worship is missing exactly the thing that makes preaching preaching — what Wax calls “the heart.”
We’re not going to argue around this. Wax is aiming at exactly the right target: handing the whole sermon over to AI. Reading the text, picturing the faces in the pews, wrestling with application, and finally speaking that word in your own voice — none of that can be outsourced, and none of it should be.
Research is not the same task as preaching
Here’s the distinction that matters. Sermon prep bundles together two very different kinds of work. One is gathering and organizing material — original-language analysis, comparing commentaries, tracing historical background, mapping a passage’s theological arc. The other is deciding what to say with it — which point to emphasize, which illustration fits, how it lands in a specific congregation’s life this week. Wax’s critique is aimed squarely at the second.
The first is the only thing Didymus Lab produces. It functions the way a seminary research assistant or a reference librarian would: “here’s what the commentaries say about this passage, here’s how the original language reads, here’s the background.” A research assistant pulling together sources doesn’t turn the resulting sermon into “the assistant’s sermon.” A report doesn’t either. What to emphasize, how to apply it, and what words to actually preach remain entirely yours.
The way pastors report actually using AI backs up that distinction. Barna Group’s 2026 survey found the top uses were idea brainstorming (50%) and Bible/theological research (36%) — while only 24% said they used AI directly for “sermon writing or editing” (up from 12% in 2024). Most pastors, without anyone telling them to, are already treating AI as a research assistant rather than a ghostwriter.
A Korean seminary drew the same line
This distinction isn’t a novel workaround — it shows up independently elsewhere. Chongshin University, a Korean Reformed seminary, issued generative-AI guidelines in 2025 built on the premise that “AI, as a tool used by humans made in God’s image, must remain under human ethical control and guidance.” Its policy restricts AI use to a supporting role — idea exploration, source lookup, grammar checking. A tool is a tool. The problem isn’t that the tool exists; it’s the moment a tool tries to sit in the preacher’s chair.
How Didymus Lab actually holds that line
That boundary doesn’t hold on good intentions alone. A Didymus Lab report never produces a “finished sermon” as an output in the first place. It’s built out of reference sections — passage overview, background, original-language notes, commentary comparison — that a pastor opens selectively, and nowhere in it does it state a conclusion or tell you what to apply. What to emphasize, which illustration to reach for, how it lands for this specific congregation — all of that stays entirely unwritten, left for the pastor to fill in. The ghostwriting risk Wax names structurally can’t happen here, because there’s no finished product to ghostwrite in the first place.
Next: how that line actually holds in practice — why a draft is still a draft, and why the final review stays a human job.
Comments
Leave a comment
Comments are reviewed before being published. Only your name and comment are stored — no personal data collected.
You might also like
Sermon Prep
The Hours You Used to Spend in Commentaries — Where Do They Go Now?
AI's real value in sermon prep isn't finding more sources — it's the harder job of narrowing them down. Here's why more material usually means more prep time, not less, and why selection, not search, is the thing actually worth automating.
Sermon Prep
Could This Report Be Quietly Feeding Me Something Theologically Off?
The deepest fear about AI in sermon prep isn't hallucination or ghostwriting — it's doctrinal drift you might not catch. Here's how Korean denominations and seminaries have already staked out public positions on AI, and how source restriction and doctrinal gating work together to guard against it.
Sermon Prep
The Last Read-Through Is Still Yours
Research shows that AI-drafted, human-reviewed workflows outperform either extreme alone. Here's why a Didymus Lab report is built as reviewable reference material rather than a finished sermon manuscript — and how to actually work through one.
Loading comments…