An AI headshot is a professional-style portrait produced by an image model from one or more existing photos of you, rather than photographed in a studio. The output is a new image — not a filtered or retouched version of the original.
How it works
Two approaches dominate:
Training-based. You upload 10–30 varied photos. The service fine-tunes a model on your face over several hours, then generates portraits from it. More input, longer wait, many variations.
Reference-based. You upload a single photo. The model uses it directly as a visual reference, generating in about a minute. Less input, faster, fewer variations per run.
What it is not
An AI headshot is not a retouched photo. Retouching modifies pixels in a photo you took — smoothing skin, adjusting light. An AI headshot generates an entirely new image in which your clothing, background, lighting, and pose may all differ from any photo you own.
That distinction matters for expectation-setting: you are not getting your selfie improved, you are getting a new portrait of your face.
The quality question that matters
The useful measure is identity retention — does the output still look like you to someone who knows you? Models can drift toward a generic attractive face, which is precisely what a colleague or hiring manager notices as "off."
Polish is easy. Recognisability is the hard part, and it is what to judge on.
Is using one acceptable?
For professional profiles, the emerging norm is that an AI headshot is fine when it represents you accurately — same approximate age, build, and appearance. The line people object to is misrepresentation, which is the same line that applied to heavy retouching before AI existed.
Related: what is a headshot pack, AI photo editor vs generator, and headshots by profession.