Can AI write a sermon?
Two years ago, 13% of church leaders said they totally opposed using AI in church work. Today it is 2.4%. The 2026 State of AI in the Church Survey — 509 pastors and church leaders, conducted by Exponential AI NEXT, AIforChurchLeaders and ChurchTechToday — found that 93.5% are actively engaging with or exploring AI, 43% use it every day, and 64% use it for sermon research or writing. “AI has moved from novelty to normal,” the survey says. The question is no longer whether preachers use AI. It is what using it does to them.
That is the question Collin Hansen, editor of The Gospel Coalition, put in his essay “Why AI Will Write Your Next Sermon.” “Does AI mean that we don't have to study and remember the Bible as much?” he asked. “If we do that, it turns what used to be Bible study into some kind of science in which you can chop it up into pieces and then process it somehow.”
Can you still study without it?
The survey's numbers describe the engaged end of the church — leaders interested enough in AI to take a survey about it. Lifeway Research's broader survey of more than 1,000 Protestant pastors found fewer than half using AI at all. But among the engaged, the habit is now daily, and the worry Hansen names is dependence. As Terry Mattingly puts it in his column on the survey, “The issue is whether many clergy can flip a switch and return to life without AI.”
The computer, Hansen says, “doesn't recognize what happens in your counseling, and in your evangelism, and in your apologetics, and in your death-bed ministry when there is a difference between asking AI to come in and supply the answers you need, instead of your own personal, intimate familiarity with the scriptures.” His challenge to pastors is blunt: “Efficient for WHAT? What is this THING that you are trying to accomplish? What are you going to say to these people? How are you going to pray for them if you are not actually formed in the scriptures and in prayer yourself?”
Whose ethics?
Hansen's second worry is sharper. “Does AI know better how to answer Bible questions than a seminarian or your average pastor? Absolutely, it does,” he says. “But if you turn around and say, 'But are you going to get good, ethical, trusted advice from AI?' Well, no, not unless you ASK for it.” Ask about the Ten Commandments and “they will hear the 'mean ethical take of an AI engineer in Silicon Valley in the year 2026, a human pretending to be an objective computer. … They will not hear the God who thundered on Mount Sinai.”
That is the same worry the survey's leaders name first: 75% call theological misalignment their top ethical concern — AI-generated content that sounds authoritative while being doctrinally wrong. It does not know your statement of faith. It does not know your tradition. It generates the most statistically likely answer, not the most theologically faithful one.
The honesty gap
Then there is the question the congregation is asking. In Lifeway's survey, 62% of pastors worry that users are not being honest about how they use AI in ministry, and 55% agreed that “God has always shared His Word through people, and AI isn't a person.” The State of AI survey found that 51% of churches have not addressed AI use with their congregation in any way at all, and only 7% have a formal disclosure statement — while 87% of leaders say their people are already asking questions about it.
The point is not that every use needs announcing. Research help is research help — you do not disclose a concordance. The point is that the congregation should never have to guess whether the words they heard were yours. The site's disclosure template has a line for exactly this: “Sermons, lessons, and prayers are written by the person delivering them. AI does not write what is spoken, sung, or prayed in our services.”
The line Rome and Salzburg draw
Two weeks ago in Rome, the Builders AI Forum — church leaders, developers and policymakers at the Pontifical Gregorian University, 5–6 November — put the same question in institutional terms. Its founder, Matthew Harvey Sanders, told the gathering: “A machine that can be blamed is a convenient place for everyone above it to leave the blame. So before any parish, school or hospital lets an agent act for a person, someone must be able to stop it, and someone must answer for it.” Cardinal Michael Czerny, contrasting Babel and Jerusalem, asked: “Out of what desire shall we build, deploy and employ artificial intelligence?” Pope Leo XIV, in Magnifica Humanitas, called AI “a valuable tool” that “calls for a measured and vigilant approach.”
Days later, the Austrian bishops drew the same line in binding form: “The use of AI in the celebration of the sacraments, preaching, liturgy, or pastoral care may only be supportive. Replacing human relationships, pastoral conversations, community, or rituals with AI is inadmissible and is expressly rejected.” In preaching above all, they said, “the irreplaceability of personal encounter and human relationship must be preserved.”
What to do this week
So can AI write a sermon? It can produce one. The question is whether the preacher who delivers it studied the passage, prayed over it, and can stand behind every word — and whether the congregation knows the answer. Three things, this week:
- Study first, AI second. Write from your own reading of the passage; then let AI critique the outline, not supply it. The practical line — where AI helps and where it stops — is in the sermon piece.
- Say what you use. One line in your disclosure, and the congregation never has to guess.
- Write the policy. Only 9% of churches have one. The weekend template takes a weekend; the three additions from the v2 piece take thirty minutes.
The tools are not going away — 2.4% opposition says that plainly. The question Czerny asked in Rome is the one to carry into the study: out of what desire shall we use them? If the answer is to preach better, the tools can help. If the answer is to preach without studying, no tool can save the sermon.
Sources: Terry Mattingly, “Can AI write a sermon? Pastors weigh the risks,” Times Record News, 21 November 2026 — the State of AI in the Church Survey figures (93.5% engaging or exploring; 43% daily, double the 2024 rate; 78% weekly or more; 64% sermon research or writing; 9% with a formal policy; opposition down from 13% to 2.4%), the Lifeway figures (fewer than half of Protestant pastors using AI; 62% worried about honesty; 55% “God has always shared His Word through people, and AI isn't a person”), and the Hansen quotations. Exponential AI NEXT / ChurchTechToday, The 2026 State of AI in the Church Survey Report — 509 pastors and church leaders; the 75% theological-misalignment and 67% human-connection figures, the 51% who have not addressed AI use with their congregation, and the 7% with a formal disclosure statement are as reported by Exponential's Kenny Jahng. National Catholic Reporter, “Rome forum explores how AI can serve Catholicism, guided by Pope Leo's warnings,” 20 November 2026 — the Builders AI Forum (5–6 November, Pontifical Gregorian University), the Sanders and Czerny quotations, and Pope Leo XIV's “measured and vigilant approach” (Magnifica Humanitas). The Austrian quotations are from the bishops' conference statement as quoted in the Austria article. One limit: the State of AI survey draws on leaders engaged enough to take a survey about AI, so its figures describe the engaged end of the church, not all pastors — Lifeway's broader survey is the check on that. The 64% sermon figure and the practical how-to are covered in the sermon piece; this piece takes up the dependence, honesty and accountability questions the new reports raise.
Related reading: Writing a sermon with AI without losing your voice — the practical line, where AI helps and where it stops; Who is speaking — you or the algorithm? — the same question from the priestly-formation side; A one-page AI disclosure your church can copy — the disclosure line quoted above; A church AI policy in a weekend — the copy-ready template; Austria just wrote its first binding national church AI rules — the binding framework quoted above; Responsible AI — the site's guide to privacy, accuracy and trust.
Researched and drafted with the help of Muse, an AI assistant. Every claim is checked against its sources, and every article is approved by Aris Angelakopoulos before publication.