What to tell wedding clients who ask about AI¶
It happens at the consultation, usually toward the end, right after the timeline questions and right before the deposit talk. The bride leans forward. "One more thing. You're not going to, like, feed our photos into an AI, are you?"
Three years ago you would not have heard that question at a wedding consultation. A year ago you might have heard it once a season. This year, some photographers are hearing it at every other meeting.
The couple is not being paranoid. They have watched Meta tell its users their public Instagram photos would train a commercial AI model. They have read about actors and voice artists suing over likeness cloning, and watched SAG-AFTRA stay on strike for nearly a year over AI digital-replica protections. They know the phrase "training data" and they have a fuzzy but real sense that their wedding photos could end up somewhere they did not agree to.
And most of us, when the question lands, do not have a clean answer ready.
At a glance
- Couples are increasingly asking wedding photographers about AI at the consultation table
- Four common places a photo can leak into AI training: cloud storage, delivery platforms, AI culling tools, metadata tools
- A one-breath script names each tool in the chain and what it does with the photos
- Jade GT is the metadata-step link: browser-only, no upload, no AI training
The real question is about trust, not technology¶
When a couple asks about AI, they are not asking you to explain transformer architectures. They are asking a simple trust question in borrowed technical language: where do the pictures of our wedding go, and who gets to look at them after we do?
You can answer that. You do not need a machine-learning degree to answer it. You need to know what your own tools do with the files you hand them, and you need a short, honest way of saying it out loud.
This post is a pocket script for the next time it comes up.
Where can a wedding photo quietly end up training an AI?¶
Before you can answer the couple, you need to audit your own pipeline. Here are the four common leak points, in order of how often photographers overlook them.
| Leak point | Common tools | First audit question |
|---|---|---|
| Cloud storage sync | Dropbox, Google Drive, iCloud, OneDrive | Does my plan's default opt me in to AI training? |
| Delivery platform AI features | Pic-Time, Pixieset, ShootProof | Is the AI feature processed on-device or in the cloud? |
| AI culling tools | Aftershoot, Narrative, Imagen | Does this tool train on my clients' photos? |
| Third-party metadata and editing | Batch renamers, geotaggers, compressors | Does this tool process in-browser, locally, or upload? |
1. Cloud storage sync, set to default¶
Dropbox, Google Drive, iCloud, OneDrive. Dropbox found this out publicly in 2023, when a default-on toggle started routing user file contents to OpenAI. The feature was later revised, but the pattern keeps reappearing across providers. Some plans train on user content by default unless you dig into settings and turn it off. Some do not. Both the policies and the defaults change. If your delivery backup runs on a consumer cloud plan and you have not read the terms recently, you do not actually know what is happening to the files.
What to do: Read the current AI and training clause for whatever cloud you use. Screenshot the setting. Revisit it every six months.
2. Gallery and delivery platforms with "AI features"¶
Several delivery platforms have added AI features in the last year: auto-culling, smart search, face clustering, style transfer. Some process on-device. Some process in the cloud. Some retain, some do not. The photographer-facing copy is often vague about the difference.
What to do: Ask your delivery vendor, in writing, two questions. "Are my client photos used to train any AI model, yours or a third party's?" and "Where is the processing happening: on my device or on your servers?" File the answer.
3. AI culling tools¶
Aftershoot, Narrative, Imagen. The good news is these companies have been transparent about what they do. Most process locally or in dedicated environments, and they have clear statements about training. Read those statements anyway. They are typically in the privacy policy, not the marketing page.
What to do: Know which of your tools are on-device and which are cloud-processed, and be able to say which one you use in one sentence.
4. Third-party metadata and editing tools¶
Batch renamers, geotaggers, keyword taggers, compression tools. This is a long tail of small utilities most photographers use without thinking about where the files go. Some run locally. Some upload. Most do not tell you on the landing page.
What to do: For anything that touches your RAWs, know whether it processes in-browser, locally installed, or in the cloud.
What do you actually say at the consultation?¶
Once you have done the audit above, you can answer the couple's question cleanly, in roughly one breath. Here is the shape of an answer that works:
A one-breath answer that works at the consultation table
"Good question. I'm hearing it more this year. Here's what happens to your photos. I cull them in [tool], which runs [on my machine / in their cloud and doesn't train on images]. I edit in Lightroom, which stays on my computer. I geotag and add metadata in Jade GT, which runs in my browser and never uploads your files. I deliver through [platform], which stores them and doesn't train on them. Nothing I use trains a model on your wedding. If that ever changes, I'll tell you."
Notice what this does and does not do:
- It names the tools. Vague reassurance sounds like a dodge.
- It admits it is a real concern, not a paranoid one.
- It makes a promise you can keep. If a tool's policy changes, you will notify.
- It does not moralize, lecture, or bash AI broadly. Couples who use ChatGPT to write their vows do not want to be preached at about AI. They want their photos treated with care.
If the couple wants more detail, they will ask. Most of the time, this is enough and the conversation moves on.
Where Jade GT fits in this¶
I'm a photographer, but I don't shoot weddings. My father-in-law did, for years in Traverse City, Michigan, and the metadata step was the part of his Monday he complained about most. The wedding pros I've been talking to while building Jade GT describe the same leak: renaming, copyrighting, geotagging, and keywording is the link in the chain where the available tools want to upload client RAWs to a server somewhere. That is the part Jade GT replaces.
Jade GT runs entirely in your browser. Your files never leave your machine. Nothing trains an AI on your work, because the tool cannot see your files in the first place. That is the architecture, not a policy we promise to keep. The full privacy details lay out what that means in practice.
It is one link in the chain. You still have to audit the others. But it is one fewer thing you have to hedge about at the consultation. For the full workflow walkthrough of that one link (rename, copyright, geo, keywords, rating, and title on 2,000 RAWs in a single pass), here is the Monday-morning version.
What this post is not¶
Three edges to keep in mind
- Not legal advice. Your jurisdiction's contract requirements, your studio's MSA, and your insurance are outside this scope. If you need clauses about AI use for client contracts, talk to a photography-specialist attorney.
- Not a condemnation of AI tools. AI culling and AI editing tools can be excellent and ethical. The issue is not whether AI touches photos. It is whether photos train models, and whether the client consented.
- Not a one-time audit. Tool policies change. Do the audit now, and put a recurring reminder on your calendar to redo it at least twice a year.
FAQ¶
Are couples really asking this?
Yes, and increasingly. It is more common with couples who work in tech, media, law, or healthcare, but it is spreading. If you have not heard the question yet, you will.
Do I have to disclose that I use AI culling to my clients?
Ethically, most wedding pros I know now mention it once during the consultation, framed around turnaround speed. Legally, it depends on your state and your contract. The safest posture is proactive disclosure in plain language.
What if a vendor's privacy policy is ambiguous?
Email their support and ask in writing. "Do you train AI models on customer-uploaded photos?" Save the reply. A clear no in writing is worth more than a paragraph of marketing copy.
Is this going to make the consultation weird?
The opposite. Couples who bring it up are relieved when you take the question seriously and have an answer ready. The awkward version is the one where you fumble.
Have the answer ready before you need it¶
Most photography-business advice says: raise your prices, book better clients, deliver on time. Still true. But the trust edge is shifting, and the couples who ask this question in 2026 are the same couples who will refer you in 2027.
The answer takes five minutes to prepare and a single sentence to deliver. Worth it.
If you want to see the no-upload architecture in practice, the Jade GT user guide has the tour.
Is the AI question showing up in your consultations? I'm collecting real language wedding pros are using with clients. Reply to this post or email me directly. Future posts in this series come straight from what working wedding pros are actually asking.