Something new is happening on the Christian music charts. Songs credited to AI artists are climbing them. Some are more transparent about how they were made than others. Some sound remarkably like a human worship band. And a growing number of worship leaders are quietly wondering whether it is acceptable to include one of them in a Sunday morning setlist.
The question is not going away. It is worth thinking through carefully now, when the pressure is low, rather than the morning it lands in your rotation without warning.
The question underneath the question
At first, this looks like a music licensing question, or a technology question, or a purity question. It is all of those, but the real question is quieter. Does the origin of a song matter to what it does when sung by a congregation? For most of church history, we have assumed the answer is yes. The person, the story, the season a hymn came out of shaped how it landed. "It Is Well" carries the death of four daughters in the wake it leaves. "Amazing Grace" carries the conversion of a slave trader.
An AI song has no such wake. It has patterns of language pulled from millions of past worship songs. It can be exquisite. It cannot be witness.
Why churches are still tempted
The temptation is real, and not shameful. The best AI worship songs are already better than the average submission a worship leader receives from a well-meaning member. They rhyme cleanly. They fit under a congregation's voice. They come with the option of any key. They can be generated to fit the exact theme of this week's sermon in under an hour.
For a small worship team pulling together a set in the middle of a busy week, that is a genuine gift. And that is precisely why the decision needs to be made before it is made in a hurry.
A frame that has helped churches we work with
We are not the ones to prescribe a policy for your congregation. But we have watched churches sort through this with a few questions that hold up.
Would you be comfortable telling your congregation, from the front, that the song they are about to sing was generated by AI? If yes, you are probably on solid ground. If no, that discomfort is worth listening to. A congregation singing something under a false impression of its origin is not worshiping the same way it would if it knew.
Does the song carry doctrine you can verify? AI-generated lyrics can subtly drift into universalism, into prosperity language, into whatever the model absorbed from its training data. Every AI song used in a service, if you decide to use one at all, needs the same lyric review your original submissions get. Probably a more careful one.
Whose name is attached to it, and what happens to money paid for it? If a real songwriter fine-tuned a model with their own work, credited themselves, and receives the licensing revenue, that is a different situation than a fully synthetic song attributed to a fictional artist whose earnings route somewhere opaque. Faith-based congregations giving to music they use should be able to trace where that giving lands.
What we think most churches should do this year
We think most churches will settle in one of three places. Some will decide not to use AI-generated worship music in services at all, and that is a defensible line. Some will use it only in explicitly labeled, non-congregational settings: background music before service, video reels, meditation clips. Some will use it in services with full disclosure, treating it as a co-write between a human worship pastor and a tool, credited that way.
What we would gently discourage is drifting into it without a decision. The churches that will regret this in five years are not the ones with strict policies or generous ones. They are the ones that sang an AI song without knowing it, discovered later, and had to explain to their congregation why nobody had said anything.
The pastoral side of the answer
Worship music is not just content. It is the language a congregation uses to speak to God together. Language shaped by a real human wrestling with grace, doubt, and joy carries something no model can replicate. That is not a technology critique. It is a theology of witness.
The question is not whether the AI song sounds worshipful. Many of them do. The question is what your congregation is being formed by over years of Sunday mornings, and whether the shepherds of that congregation are still choosing the diet, or whether an algorithm is starting to.
What to write down
A short paragraph in your worship team handbook is probably enough. What you will and will not consider. Who makes the call when a case is unclear. Whether songs generated by AI will be disclosed to the congregation, and how. Any licensing platforms you will and will not source from.
Three or four sentences. Reviewed once a year. That is more thought than most churches have given this question so far, and it will save you a hard conversation later.
The wider tech decision
This is one of the first of many decisions like it that churches are going to face this decade. AI-generated devotionals. AI voice cloning of past pastors for archive material. AI-composed prayers included in bulletins. Each one will feel a little different, and each one deserves the same slow question about origin, disclosure, and what a congregation is being invited to trust.
Your church does not need to have all of those answers now. What it needs is a habit of asking the question out loud, together, before a piece of technology becomes normal without anyone having chosen it.
If you would like a hand thinking through your church's approach to AI in worship, communications, or the website, and shaping small documents that make those decisions easier for the volunteers who carry them out, reach out through our contact page and tell us where you are starting from.