If you’ve been experimenting with AI image tools — Midjourney, DALL-E, Adobe Firefly, or one of the many others — you’ve probably noticed they can produce surprisingly decent visuals in seconds. Need a stock-photo-style image of people gathered around a table? A colorful background for your worship slides? An illustration for a social media post about your community garden project? These tools can generate something passable before you finish your coffee.
But “passable” and “appropriate for church communications” aren’t always the same thing. AI-generated images raise real questions about authenticity, representation, and taste that are worth thinking through before you start dropping them into your bulletin or your Instagram feed. Here’s a practical guide to what works, what doesn’t, and where the ethical lines are.
The short version
- AI image generators are genuinely useful for certain church communication tasks — backgrounds, textures, abstract illustrations, and placeholder images during the design process.
- They can get weird fast when you ask for realistic people, especially in ministry contexts where authenticity and representation matter deeply.
- There are real ethical concerns around representation, consent (or the lack of it in training data), and the difference between illustration and deception.
- A simple disclosure practice — telling your audience when an image is AI-generated — goes a long way toward maintaining trust.
- Bottom line: AI images work best as a supplement to real photography of your actual community, not a replacement for it. Make sure your church has a photo team always capturing real photos during events and fellowship. Load the photos into a common source (like Smugmug) to make them readily accessible to your communications team.
Where AI images actually help
Let’s start with what works. AI image tools are genuinely good at a few things that church communicators regularly need:
Abstract and decorative visuals. If you need a watercolor-style background for your sermon series slides, a textured gradient for a social media post, or an abstract illustration of something conceptual like “community” or “hope,” AI tools handle this well. These images don’t pretend to be photographs of real people or places, so the authenticity question doesn’t come up.
Quick mockups and placeholders. If you’re designing a new page for your website and you need a placeholder image to show the layout before you schedule a photo shoot, AI-generated images work fine as temporary stand-ins. Just make sure you replace them with real photos before the page goes live.
Illustrative graphics for teaching. If you’re creating a Bible study handout and you want a simple illustration of a first-century marketplace or a map-style graphic of Paul’s missionary journeys, AI tools can produce something more engaging than clip art without the cost of hiring an illustrator. The key here is that no one expects these to be photographs — they’re clearly illustrations serving an educational purpose.
Event promotion backgrounds. Need a fall-themed background for your harvest festival flyer? A warm, inviting texture for your Christmas Eve service invitation? AI tools produce these quickly and they look professional. Again, these work because they’re decorative, not representational.
Where it gets weird
The trouble starts when AI images try to depict real-seeming people in ministry contexts. And this is exactly where most church communicators are tempted to use them — because real photography of diverse, welcoming congregational life is expensive, time-consuming, and hard to get right.
Fake congregation photos. If you generate an image of smiling people sitting in pews and use it on your “Visit Us” page, you’re showing visitors a congregation that doesn’t exist. This might seem like a small thing, but think about what it communicates: “Come visit us — here’s what you’ll find!” followed by a completely fabricated image. If a visitor shows up expecting the diverse, well-dressed, perfectly lit group they saw on the website and finds something different, you’ve started the relationship with a misrepresentation.
The uncanny valley of AI faces. AI-generated people often look almost right but not quite. Hands with too many fingers have become a running joke, but the subtler issues — slightly off skin textures, eyes that don’t focus on the same point, teeth that are uniformly perfect — create an eerie quality that viewers notice even if they can’t articulate it. In print at small sizes, you might get away with it. On a screen at full resolution, it reads as off.
Representation without reality. This is the big one. AI image tools make it trivially easy to generate images showing racial, age, and ability diversity. And for churches that are working toward greater diversity, the temptation is real: you can show the congregation you aspire to be, rather than the one you currently are. But this crosses a line. If your church is predominantly one demographic and your website imagery shows something different, that’s not aspiration — it’s false advertising. Visitors from underrepresented groups who show up based on that imagery will feel deceived, not welcomed.
The ethical questions worth sitting with
Beyond the practical concerns, there are deeper ethical issues that deserve attention in faith communities especially:
Training data and consent. Most AI image generators were trained on millions of images scraped from the internet — photographs taken by real photographers, of real people, often without their knowledge or consent. When you generate an image that resembles a particular style, you’re building on creative work that the original artists were never compensated for. Some tools (like Adobe Firefly) have made efforts to use only licensed training data, which is worth considering if this matters to your community.
The labor question. Every AI-generated image is one that a real photographer, designer, or illustrator wasn’t hired to create. For large organizations with design budgets, this might mean efficiency. For the freelance photographer in your congregation who does the church directory photos, it might mean lost income. This isn’t a reason never to use AI images, but it’s worth being thoughtful about — especially for faith communities that talk about supporting workers and fair compensation.
Honesty and trust. Churches trade on trust. If a congregant or visitor discovers that the warm, inviting images on your social media were generated by a computer rather than taken at actual church events, it can feel like a betrayal — even if the images were innocuous. The reaction isn’t “those are nice graphics.” It’s “what else here isn’t real?”
A practical framework for deciding
Here’s a simple set of questions to run through before using an AI-generated image in any church communication:
Does this image depict real-seeming people? If yes, use a real photograph instead. If the image is abstract, decorative, or clearly illustrative, AI generation is generally fine.
Could a viewer reasonably believe this is a photograph of our community? If yes, don’t use it — or label it clearly as an illustration. The test isn’t your intention; it’s the viewer’s reasonable interpretation.
Am I using this because we don’t have real photos? If so, the better investment is organizing a photo day. Ask a skilled photographer in your congregation (there almost certainly is one) to spend a Sunday morning capturing real moments. Those images will serve you better for years than any AI-generated alternative.
Would I be comfortable if a congregant asked, “Is this photo real?” If the honest answer would make you squirm, choose a different image.
Am I disclosing that this is AI-generated? A simple note — “Illustration created with AI” in small text, or a mention in your social media caption — maintains trust and normalizes transparency. It costs nothing and it signals that your church values honesty in small things as well as large ones.
What to watch out for
Don’t use AI images of children. Even in illustration form, AI-generated images of children in church contexts carry risks around safeguarding perception that aren’t worth the convenience. Use real (consented, release-signed) photos of your children’s ministry, or use graphics without people.
Watch for cultural and religious stereotypes. AI tools tend to reproduce stereotypes from their training data. If you ask for an image of “a pastor,” you’ll likely get a white man in a suit. If you ask for “a church,” you’ll get something that looks more like a New England steeple than the storefront congregation down the street. These defaults reflect bias, not reality, and reproducing them in your communications reinforces the problem.
Don’t mix AI and real photos without distinction. If your website uses real photographs of your congregation on some pages and AI-generated images on others, and there’s no visual distinction, you’re creating an inconsistency that erodes trust. Pick a lane for each context and be consistent.
Check the terms of service. Not all AI image tools grant the same rights to the images you generate. Some restrict commercial use, some retain rights to the images, and the legal landscape is still evolving. For church use (which is usually non-commercial), this is rarely a problem, but it’s worth a quick check — especially if you’re using a free tier of a tool.
A small next step
This week, do a quick audit of the images on your church website and your most recent social media posts. For each image, ask yourself: is this a real photo of our community, a stock photo, or something generated? If any AI-generated images are standing in where real photos should be — especially images of people — flag them for replacement. And for future use, adopt a simple rule: AI for decoration and illustration, real photos for people and places. Add a brief disclosure line to AI-generated graphics, and you’ll stay on the right side of trust.
If posts like this are useful, you can get them in your inbox — subscribe to the Faithlab blog. No sales pitches, just practical help for church and nonprofit communicators.