Guest post
Every wildlife photographer knows the feeling. You spend three dawns in a hide, finally get the shot, post it, and within an hour someone asks whether it is AI. A few years ago that question was an insult. Today it is a fair one. Image generators can produce a snow leopard on a ridge at golden hour in seconds, and photo competitions, stock libraries and editors are all dealing with the fallout.
This guide is for photographers on both sides of that problem: protecting your own work from suspicion, and checking images you are asked to judge, license or share.
Why wildlife and nature images are such an easy target
Generators are trained on millions of nature photographs, so they are very good at the look: shallow depth of field, creamy bokeh, perfect rim light. What they are still weak at is the biology and the physics. An animal photographed in the field obeys rules a generator only approximates. That gap is where most fakes give themselves away.
Six things to check by eye
1. Anatomy that does not add up
Count toes, claws and teeth. Look at where feathers meet the beak and where fur meets the eye. Generated animals often have an extra digit, a beak that blends into the face, or whiskers that start from nowhere.
2. Fur and feather texture that is too uniform
Real fur clumps, gets wet, catches dust and changes direction around muscle. Generated fur tends to flow in one smooth, painterly direction across the whole body.
3. Catchlights and reflections
Both eyes should show the same light source in roughly the same position. Water reflections should mirror the animal correctly, including the underside. Generators frequently get one eye right and the other wrong, or reflect a slightly different animal.
4. Backgrounds that melt
Look past the subject. Branches that end mid-air, grass that turns into texture, or a horizon that bends behind the animal are common tells.
5. Behaviour that is too perfect
Experienced photographers know what an animal actually does at that time of day, in that season, in that habitat. A barn owl hunting in full midday sun over a tropical beach should raise a question even if every feather looks right.
6. Missing camera data
A raw file or an original JPEG from a camera carries EXIF data: body, lens, shutter speed, aperture, ISO, often GPS. A generated image usually has none, or carries the name of the tool that made it.
Use a detector as a second opinion
Eyes are good at catching obvious fakes. They are much worse at catching good ones, and they are worse still when you are looking at a small image on a phone. That is where a dedicated checker helps.
A practical option is this free deepfake image detector, which needs no account for a quick check. It gives a verdict and an AI score for the image, and it runs a separate source check that reads the file's metadata and Content Credentials and, when the file still carries that evidence, names the generator, such as Midjourney or DALL-E. That second part matters for photographers: a screenshot or a re-saved copy loses its metadata, so always test the original file you were sent.
Treat any detector result as evidence, not a final ruling. Combine it with the visual checks above and, for anything that matters, ask for the raw file.
Protecting your own work
Keep your raws. A raw file with intact EXIF is the strongest proof you were there.
Shoot a sequence. A burst of frames showing the animal moving is very hard to fake convincingly.
Turn on Content Credentials where your software supports them. Several cameras and editors can now attach a signed record of how an image was captured and edited.
Be open about edits. Sky replacement, heavy cloning or composites are fine in the right context, but say so. Hidden edits are what make audiences suspicious of everything else.
For competition judges and editors
Ask for raw files from shortlisted entrants as a standard step, not an accusation. Run finalists through a checker and the eye test, and look at the entrant's other work from the same trip. A single perfect frame with no sequence around it deserves a closer look.
The bottom line
AI images are not going away, and neither is the doubt they create. The photographers who come out ahead will be the ones who can prove their work is real, and who know how to check when someone else's might not be.
If you’d like to order a framed print of one of my wildlife photographs, please visit the Prints page.
If you’d like to book a lesson or order an online photography course, please visit my Lessons and Courses pages.
