By Brent Lacy

Worship is where the church’s theology becomes embodied. The songs you sing, the prayers you pray, the liturgy you follow, these are not decorative. They form your congregation’s understanding of God, shape their emotional and spiritual vocabulary, and carry the weight of the church’s gathered life before God. A congregation that sings shallow songs for twenty years will have a shallow theology. A congregation that prays rich, honest prayers will have a richer prayer life. The content of worship matters more than most churches realize.

So when a worship leader asks whether they can use AI to help plan Sunday’s service, the question is not just practical. It is theological. And the answer is not simple, because the question is not simple.

The Barna/Gloo 2026 research found that pastors are using AI primarily for “behind-the-scenes work, as a thought partner, a visual aid.” That instinct is sound in worship contexts too. The challenge is that the line between preparation and replacement is easier to cross in worship than almost anywhere else in ministry, because worship involves so many tasks that feel administrative but are actually formative.

What AI Can and Cannot Do in Worship

AI tools can generate lyrics, suggest chord progressions, draft liturgical prayers, produce bulletin copy, recommend songs based on a given theme, and create visual content for worship slides. These are real capabilities, and they can save a volunteer worship leader significant time. A worship leader who spends three hours every week searching for songs that fit the sermon theme can get that down to thirty minutes with the right AI tools. That is a real benefit.

What AI cannot do is worship. It cannot bring the weight of genuine prayer to a written prayer. It cannot sense that this congregation, on this Sunday, needs a particular song because of what they have been through this week. It cannot discern the movement of the Spirit in a worship service. It cannot know that the family in the fourth row just lost their father, or that the teenager in the back is asking questions about faith for the first time, or that the congregation needs to lament before it can celebrate.

These are not limitations that better AI will eventually overcome. They are categorical differences between a tool and a person made in the image of God. The Christian Curator’s overview of AI and the church notes that “the incarnational nature of Christ’s ministry” is a key theological guardrail for AI in pastoral and worship contexts. Worship is incarnational. It happens in bodies, in a specific place, among specific people. AI can assist the preparation for that incarnational act. It cannot perform it.

AI and Worship Music: Where the Lines Are

Song Selection and Planning

Using AI to research songs that fit a given theme, identify songs your congregation may not know, or organize a multi-week worship plan is a legitimate and helpful use of the technology. This is research and organization, tasks where AI adds efficiency without replacing the worship leader’s discernment. You still decide which songs fit your congregation’s voice, your theological tradition, and the specific moment you are planning for. AI gives you more options to choose from. The choosing is still yours.

A practical approach: give the AI your sermon text, your congregation’s general style, and any specific constraints, songs you have sung recently, songs that are too difficult for your musicians, and ask for ten song suggestions with brief explanations of why each fits. Evaluate the list, pick the best two or three, and add your own instincts about what your congregation needs this week. The AI does the research. You do the discernment.

Original Worship Songs

AI can generate worship lyrics. The results are typically generic, theologically adequate but lacking the specificity, depth, and Spirit-led creativity that makes a worship song genuinely formative. If you use AI to generate a starting point for an original song, treat it as a rough draft that requires significant human reworking, not a finished product. The congregation that sings a song shaped by their own worship leader’s prayer and study is receiving something different from the congregation that sings a song generated by an algorithm.

There is also a copyright question worth noting. The legal status of AI-generated creative content is still being worked out in courts and legislatures. Using AI-generated lyrics in a published or widely distributed context carries legal uncertainty that does not exist with human-authored content. For a small church using a song only in its own services, this is a lower-stakes question. For a church that publishes or distributes worship content, it is worth taking seriously before you invest in developing AI-generated material.

Liturgy and Written Prayers

AI can draft liturgical prayers, calls to worship, benedictions, and responsive readings. These are appropriate uses, with the same caveat that applies everywhere: a human must review, edit, and own the final content. A pastor who prays an AI-generated prayer with genuine intention and theological care is doing something different from a pastor who reads an AI-generated prayer without engaging it personally. The difference is not in the words. It is in the person praying them.

The TRUST framework from AI and Faith researchers (McGrath, Huizinga, Dyer, and Graves, 2025) emphasizes that AI should “support rather than replace authentic human relationships, particularly relating to pastoral care and spiritual formation.” Worship is spiritual formation. The liturgy your congregation prays shapes who they become. That shaping work requires human engagement, not just human review.

Sermon-Integrated Worship Planning

One of the most useful applications of AI in worship planning is helping the worship leader understand the theological content of the sermon so they can plan music and liturgy that reinforces it. A worship leader who receives the sermon text on Monday and asks an AI tool to summarize the key theological themes, suggest related Scripture passages, and identify emotional tones the service might need to address is using AI to serve the congregation’s formation. The AI is doing research. The worship leader is doing discernment.

Specific Risks in Small Church Worship Contexts

The Generic Worship Risk

Small churches have a distinctive worship culture. The songs that mean something to your congregation, the way you pray, the rhythm of your liturgy, these are not generic. They are specific to your people and your place. AI-generated worship content tends toward the generic, because it is trained on the average of everything rather than the specific of your congregation. A worship service that feels like it could have happened anywhere is a worship service that has lost something important.

The guardrail: always adapt AI-generated content to your specific congregation before using it. A call to worship that AI generates for “a small evangelical church” needs to be rewritten to sound like it came from your pastor, in your voice, for your people. If you cannot do that rewriting, the AI-generated content is not ready to use.

The Preparation Displacement Risk

The most subtle risk in worship planning is that AI tools will displace the prayer and preparation that make worship leadership meaningful. A worship leader who used to spend an hour in prayer and Scripture before planning Sunday’s service may find that AI makes the planning so efficient that the prayer gets crowded out. The tool that was meant to free up time for prayer can quietly become a substitute for it.

The guardrail: decide before you open the AI tool how much time you will spend in prayer and Scripture first. Make that commitment non-negotiable. The AI tool is for after the prayer, not instead of it.

The Theological Accuracy Risk

AI-generated liturgical content can be theologically imprecise in ways that are hard to catch if you are not looking carefully. A call to worship that sounds beautiful but subtly misrepresents the character of God, a prayer that uses language that is theologically problematic in your tradition, a responsive reading that conflates two different theological concepts, these are real risks. Every piece of AI-generated worship content must be reviewed by someone with theological knowledge before it reaches your congregation.

Practical Guidelines for Small Church Worship Teams

• Use AI for research and administration, not for the act of worship itself. Song research, bulletin drafting, and thematic planning are good uses. Spontaneous prayer and Spirit-led song selection are not.
• Review everything before it reaches the congregation. AI-generated liturgy, prayers, and song lyrics must be reviewed by someone with theological knowledge before use in a worship service. The person who generated the content should not be the only person who reviewed it.
• Adapt everything to your congregation. Generic AI-generated content needs to be rewritten to sound like it came from your church, in your voice, for your people. If you cannot do that rewriting, the content is not ready to use.
• Be honest with your worship team about what you are using. If you use AI to help plan the service, your team should know. This is not a confession. It is normal transparency about your process.
• Do not let AI replace your prayer preparation. The most important thing a worship leader brings to Sunday morning is not a well-organized set list. It is a heart that has been in prayer. AI cannot do that for you, and no efficiency gain is worth trading it away.

A Word About Worship Slides and Visual Content

AI image generators can create visual content for worship slides quickly and cheaply. This is a legitimate use, with the same caveats that apply to all AI-generated content. Review every image before it appears in a worship service. AI image generators can produce content that is stylistically inappropriate, culturally insensitive, or theologically problematic in ways that are not immediately obvious. A volunteer who generates worship slide backgrounds without reviewing them carefully is taking a risk that is not worth taking.

A practical approach: use AI image generators for abstract backgrounds and textures, where the risk of problematic content is lower. For images that depict people, places, or specific theological concepts, use stock photography from reputable sources where you can see exactly what you are getting before you use it.

The Deeper Question

Behind all the practical questions about AI and worship is a deeper theological question: what is worship for? If worship is primarily a performance, a production to be improved for quality and efficiency, then AI is a natural fit. If worship is primarily an encounter between a congregation and the living God, mediated by human beings who have been formed by prayer and Scripture and genuine relationship with their people, then AI is a tool that must be kept firmly in its place.

Small churches tend to have a clearer answer to that question than large churches do, because small churches cannot hide behind production quality. When your worship team is three people and a piano, the encounter has to be real. The authenticity is not optional. That is not a weakness. It is a theological advantage. AI, used wisely, can help you steward that advantage. Used carelessly, it can erode it.

---

Frequently Asked Questions

Can AI write worship songs for our church?

AI can generate lyrics and suggest chord progressions, but the result is typically generic. It lacks the theological depth, congregational specificity, and Spirit-led creativity that makes worship music genuinely formative. AI is better used as a brainstorming tool than as a songwriter. Use it to generate starting points, then do the real work of shaping those starting points into something that belongs to your congregation.

Is it ethical to use AI-generated prayers in a worship service?

Using AI to help draft a written prayer, which a pastor then reviews, edits, and prays with genuine intention, is ethically defensible. Presenting an AI-generated prayer as spontaneous or personally composed without disclosure raises integrity concerns. The key is human ownership of the final content and genuine engagement with what you are praying.

What are the best uses of AI for small church worship planning?

Song selection research, thematic planning for sermon series, bulletin and order of service drafting, and generating discussion questions for worship team meetings. The rule of thumb: AI for the logistics, humans for the liturgy.

How do I evaluate whether an AI tool is appropriate for worship planning?

Ask: Does the tool have strong data privacy protections? Is the theological content reviewed by qualified humans before it reaches your congregation? Does using this tool free up more time for prayer and preparation, or does it replace prayer and preparation? If your congregation knew you were using this tool, would they feel that something important had been lost?

---

What Worship Reveals About a Church’s Theology of AI

How a church uses AI in worship reveals more about its theology than almost any other AI decision it makes. Worship is where the church’s deepest convictions are expressed and formed. A church that uses AI carelessly in worship is a church that has not thought carefully about what worship is for. A church that uses AI wisely in worship is a church that has.

The question is not whether AI can produce content that sounds like worship. It can. The question is whether the content of worship should be produced by a tool that has no relationship with God, no experience of prayer, no knowledge of the specific congregation it is serving, and no stake in the spiritual formation of the people who will receive it. The answer to that question, for most theologically serious churches, is no, at least not without significant human engagement and ownership.

This does not mean AI has no place in worship planning. It means AI’s place in worship planning is clearly defined: it is a research and logistics tool, not a liturgical authority. The worship leader who uses AI to find songs and draft bulletin copy, and then brings their own prayer, their own knowledge of the congregation, and their own theological discernment to the actual planning of the service, is using AI appropriately. The worship leader who lets AI plan the service and then shows up on Sunday to execute it has abdicated something important.

A Conversation Guide for Worship Teams

Here is a guide for a worship team conversation about AI. It is designed for a thirty-minute meeting and can be adapted for any size worship team, from a single volunteer to a full team.

Opening question (5 minutes): Have you used any AI tools in your worship planning or preparation? If so, what did you use them for? This question surfaces what is already happening without judgment.

Theological grounding (10 minutes): Read together the section of this article on what AI can and cannot do in worship. Discuss: what resonates with your experience? What surprises you? What questions does it raise?

Practical guidelines (10 minutes): Based on the discussion, identify three things your worship team will do with AI and three things you will not do with AI. Write them down. These become your informal worship team AI guidelines.

Closing commitment (5 minutes): Each person identifies one specific way they will use AI to free up more time for prayer and preparation in the coming month. Share those commitments with each other. Follow up at your next meeting.

This conversation does not need to be formal or lengthy. It needs to be honest. A worship team that has talked openly about AI is a worship team that will make better decisions about AI than a worship team that has never raised the question. The conversation itself is a form of leadership, and it is a form of leadership that any worship leader can provide, regardless of their technical expertise or their church’s size.

The Question of Authenticity in Worship

Behind all the practical questions about AI and worship is a deeper question about authenticity. What makes worship authentic? Is it the quality of the music? The theological precision of the liturgy? The emotional resonance of the prayers? Or is it something else, something harder to define and harder to produce?

Most theologically serious worship leaders would say that authenticity in worship has something to do with the genuine engagement of the people involved, both the leaders and the congregation. A worship service is authentic when the people leading it are genuinely present, genuinely praying, genuinely seeking God on behalf of the congregation. It is inauthentic when the leaders are going through motions, performing a role without genuine engagement.

AI creates a new way for worship to become inauthentic. Not through bad intentions, but through the gradual displacement of genuine engagement by efficient production. A worship leader who uses AI to generate the call to worship, the prayers, the song selections, and the bulletin copy, and then shows up on Sunday to execute the service, has produced something that may be technically excellent but is not genuinely theirs. The congregation may not be able to name what is missing. But they will feel it.

The guardrail against this kind of inauthenticity is not a policy. It is a practice. The practice of prayer before planning. The practice of sitting with the sermon text before asking AI to suggest songs. The practice of knowing your congregation well enough to sense what they need this week, not just what the algorithm suggests. These practices cannot be automated. They are the irreducible human contribution to worship leadership, and they are what makes the difference between a worship service that is technically adequate and one that is genuinely formative.

Protect those practices. Use AI for everything else. That is the whole framework for AI and worship, and it is simple enough to remember on a busy Thursday afternoon when you are tempted to let the algorithm plan Sunday’s service for you.