Is using an AI headshot misrepresentation?
The question worth asking is not whether software touched the photo. It is whether someone who meets you would recognise you from it.
This comes up a lot, usually phrased as a worry rather than a question. Is it dishonest? Will someone find out? Is it cheating in a way that a photographer would not be?
It is a fair thing to wonder about, and the answer is more useful than a simple yes or no.
Everyone edits. That has never been the issue
A professional headshot is not an unmediated record of your face. The photographer chose a lens, chose a distance, chose where the light came from, chose the frame from forty and threw the rest away, and then retouched it. That is the job. Nobody considers it dishonest because everyone understands it is happening.
So “was it edited” is not the dividing line, and it never has been. If it were, every professional headshot ever taken would fail the test.
The dividing line is somewhere else.
The actual test
Would someone who has only seen your photo recognise you when you walk in?
That is it. That is the whole test, and it is the one that matters practically, because that is the moment where a photo either does its job or embarrasses you.
Everything else follows from it:
Changing the room you were in does not affect recognition. A grey background instead of your kitchen changes nothing about your face.
Changing the light on your face does not affect recognition, as long as it stays a plausible light. It is the same thing a photographer does with a softbox.
Evening out skin does not affect recognition at normal strength. Skin varies week to week anyway. Removing a permanent, identifying feature is different, and you can feel the difference while you are doing it.
Changing the shape of your face fails the test. So does changing your apparent age by a decade, your body, or your hair into something you have never had.
Notice that the failures are all in the same category: they are changes to you, not changes to the photograph. That distinction is more robust than any rule about which software is allowed.
Where generated headshots sit
This is where it gets specific, and where the honest answer is less comfortable.
A generative service fits a model to a set of your photos and then produces new images. The output is not a photograph of a moment that happened. It is a synthesis, and it is usually a slightly idealised one, because the training data is full of idealised portraits and that pull shows up in the result.
That is why generated sets frequently come back with a subtly different jaw, better skin than any human has, and an age that is not quite yours. Not because anyone is being deceptive, but because that is what the process does.
Which does not make it misrepresentation automatically. Apply the test. If the result looks like you and you would be comfortable sitting down opposite someone who had seen it, it passes. If you are choosing the version that looks best precisely because it does not quite look like you, you already know the answer, and so will they.
We went through the visual tells in why AI headshots look off.
What people actually notice
Nobody audits your photo. What happens instead is much simpler: they form an expectation from it, and then they meet you, and if the two do not match they feel it before they can name it.
That mismatch does not read as “they used software”. It reads as slightly untrustworthy, which is a worse outcome than a plain photo would ever have produced. The cost of an over-processed photo is not being caught. It is a small, unfixable wobble in the first thirty seconds.
An ordinary, current, well-lit photo of your actual face beats it every time.
Where disclosure matters
Most of the time, nowhere. You are not obliged to annotate your LinkedIn photo any more than you are obliged to state which lens was used.
Two exceptions worth knowing about:
Contexts with explicit rules. Official documents, some professional registers, some regulated industries and some platforms have their own requirements about what may be altered. Those are rules, not conventions, and they are worth checking rather than guessing.
Contexts where the photo is the product. Acting and modelling submissions are the obvious ones. A casting photo that does not match the person is a wasted appointment for everybody, and the industry norms about this are strict for good reason.
What we do here, plainly
Since this is our own tool, it is fair to say where it sits.
It does not generate a face. It takes the photograph you already have and changes the conditions around it: it separates you from the background, puts a plain one behind you, balances the light, evens the skin a little, and crops properly. The pixels of your face stay your face.
That is a deliberate choice rather than a technical limit. The moment a tool starts redrawing features, it stops being able to promise the thing above.
The controls go far enough to look wrong if you push them all the way, which is also deliberate. The defaults are set where a photographer would leave them.
Run a photo through it and compare it to the original with the slider. If the result still looks like you, it is fine to use. That comparison is the whole answer.
Frequently asked questions
Is it dishonest to use an AI headshot?
Not inherently. The test is whether someone who has only seen the photo would recognise you in person. Changing the background and the light passes it; changing your face does not.
Can people tell if a headshot is AI generated?
Often, yes, from the skin texture, the hands, the ears and a certain glossiness. More importantly, people who meet you can tell when the photo did not match.
Do I have to say my photo was edited?
Generally no, for the same reason you do not have to state that a photographer retouched it. Check the rules where there are rules, such as official documents or casting submissions.
Is editing a photo the same as generating one?
No, and the difference is the useful one. Editing changes the photograph. Generating produces a new image of a person who resembles you.
What is a safe amount of retouching?
Anything you would be happy for a colleague to see next to the original. If you would not show them the before, you went too far.