.png)
A guest trips on a broken step by the side entrance. Your night manager does everything right. They check on the guest, take three photos of the step, and file the report before the shift ends.
Eight months later, a claim lands a those photos are the strongest thing you have and someone on the other side asks a question that almost nobody used to ask: how do we know these are real?
It's a fair question now. Today, anyone with a phone and a free AI tool can make a convincing picture of a lobby, a spill, or a stained ceiling in a few minutes, or quietly erase something from a real one. Those tools aren't going away, so the photos in your incident reports need to be able to answer for themselves.
Nobody used to ask if a photo was real. Now they have to prove it's real.
Why photo evidence is getting a second look
Insurers have noticed. In January 2026, the law firm Debevoise & Plimpton published a memo on AI-generated images in insurance claims. It shows how a real photo of minor water or car damage can be pushed into something much worse with AI, and it lists the ways insurers can check: detection software, asking for more photos, and sending adjusters out more often.
That's aimed at fraud, and fraud should get caught, but it also means honest photos will get a closer look, including yours. If an adjuster or the other side's lawyer runs your images through a detector, you want to already know what it's going to say.
Two different problems: generated and edited
"AI photo" covers two separate things, and it's worth keeping them apart.
A generated image was made from scratch by an AI tool. There's no real moment behind it.
An edited image started as a real photo and then got changed. Something was removed, added, or touched up. The moment happened, but the picture no longer shows it the way it was.
Both can weaken a report, in different ways. That's why LogEasy tells you which one it found.
What LogEasy does when a photo is uploaded
Every photo added to an incident report in LogEasy is checked for AI the moment it's uploaded. There's nothing to turn on, and nothing extra for your team to do.
- Generation detected. The image looks AI-made. Whoever added it has to write a note explaining why it belongs in the report before the report can be submitted for review.
- Editing detected. A real photo has been altered, and it's flagged so everyone reading the report knows.
We tested this ourselves before we shipped it. Our team used ChatGPT and Gemini to make a fake photo of a wet pool floor with a sign, and LogEasy flagged it as generated. Then we took a real, unedited photo, removed a dog's leash from it in Adobe, and uploaded that. LogEasy flagged it as edited, not generated. That's the distinction you'd want it to make.
.png)
The point isn't to accuse anyone. Most flags will have an ordinary explanation. What matters is that the explanation gets written down at the time, by the person who uploaded the photo, instead of being pieced together months later when someone asks.
The explanation gets written down at the time, not pieced together months later.
Four habits that keep your photos credible
Detection catches problems. Good habits stop you from creating them. These are worth sharing with your team.
%401x.png)
- Take the photo inside LogEasy. The in-app camera keeps evidence separate from personal camera rolls, so nobody has to prove later which phone a photo came from or whether it sat in someone's gallery for a week.
- Shoot it right away, from more than one angle. One close-up and one wide shot that shows where the hazard is in the room go a long way. It's also exactly what an insurer asks for when they want to verify a photo.
- Upload the original. Cropping, filters, or brightening a dark photo before upload can muddy the question of whether an image was changed. Let the original go into the report.
- Say what the photo shows. A line in the narrative like "Photo 2: cracked tread on the second step, taken from the parking lot side" ties the picture to the account and makes both harder to pick apart.
The record has to hold up later
Most incident reports never get looked at again. The ones that do usually get looked at closely, long after the shift, by people whose job is to find the weak spot. LogEasy is built for that moment. It's easy for your team to use in the middle of a busy night, and the record is hard to challenge when it finally gets read.
