AI-Generated Models and the New Impossible Beauty Standard

For years, advertising still had one awkward problem to deal with: an actual person. A real model comes with skin that has texture, features that are not perfectly even, a body that changes, a face that gets tired, and all the other rude interruptions that come with being alive. Brands could style that, light it, pose it, edit it and polish it within an inch of its life, but there was still a human body underneath the whole production causing trouble. AI-generated models solve that problem beautifully, if the goal is to remove reality from the room.

That is why AI-models are so attractive to brands. An AI model does not age, fluctuate, retain water, go off on a bad angle, show stress in the skin, or insist on having a body with history. It can be rebuilt again and again until every surface is smoother, every proportion neater and every detail more commercially obedient. Advertising has always idealised people. AI takes the final human brake off the system.

What appears on screen is no longer a person who has been edited. It is a synthetic body assembled from the visual cues that already get rewarded: clear skin, balanced features, controlled proportions, youth, thinness, symmetry and composure.

What AI-generated models change

The brain still reads AI-generated models as human enough to trigger comparison. It sees a face, a body, skin, posture, eye contact and expression. The comparison machinery does not politely stop and wait for a technical explanation. It gets on with the job. The problem is that the target has changed. The body being measured against no longer has to obey biology, history or ordinary human variation.

That matters because the gap quietly widens. With traditional advertising, there was usually at least some loose sense that the person in the image existed somewhere, however edited, however inaccessible, and however heavily supported by money, grooming, lighting and retouching. AI changes the terms completely. The viewer can still feel pulled into comparison, while the image itself drifts further from anything a human body could reliably be. That is a nasty little trick. The image keeps the realism needed to trigger self-evaluation, then drops the limits that used to keep the standard faintly tethered to life on earth.

There is a difference between exaggerating beauty and engineering it. A heavily retouched image still begins with one body, one face, and one set of features moving through time. AI imagery does not need that starting point. It can pull together the most desirable cues from a huge pile of existing visual patterns, smooth out the friction between them, and present the result as if it were simply a better-looking person. It is less “here is someone unusually attractive” and more “here is a body built to win”.

That distinction matters because the image starts to feel cleaner and more final. There is no sense of compromise in it. No slight asymmetry left behind. No texture escaping the clean-up job. No feature refusing to cooperate with the rest. The body becomes pure output. Once that starts showing up across campaigns, shopfronts and feeds, the standard itself shifts. The ideal is no longer a highly selected human. It becomes a synthetic performance of perfection that still wears the costume of realism.

Real people, even in polished advertising, still tend to carry traces of life: slight asymmetry, skin that behaves like skin, tiredness, tension, puffiness, and the small unpredictable details that make a face or body look human. AI can wipe those out with eerie efficiency. Skin becomes uniformly smooth, proportions start to look suspiciously neat, and expressions sit in that polished dead zone where everything looks serene and slightly vacant at the same time. The final image has the strange cleanliness of something that has never had a bad night’s sleep, never felt bloated, never had a stress breakout, and never once lost a fight with gravity.

Repeated exposure to imagery like that can shift the baseline for what starts to feel normal. Texture begins to look wrong, age begins to look like decline, and weight fluctuation begins to look like a personal failure. Human variation gets recast as evidence of falling short. That shift can happen even when people know perfectly well that the image is synthetic. Knowledge helps, but it is not a forcefield. The comparison system is quick, automatic, and often thoroughly unimpressed by rational objections.

What AI-generated models do to our mental health

AI-generated bodies also strengthen one of the most dangerous lies in appearance culture: the idea that the body is infinitely controllable. Synthetic bodies do not eat, retain water, plateau, recover slowly, get injured, deal with hormonal changes or respond unpredictably to stress. They present a body with no friction and no limit. That can feed directly into eating disorder logic, where control, optimisation and visible mastery already carry enormous emotional weight.

That same logic does not stop at eating disorders. It also lands hard on people whose distress is organised around perceived defects rather than weight or control.

Body dysmorphia already involves distorted perception, obsessive checking and a brutal fixation on perceived flaws. AI-generated models can slot neatly into that machinery and make it worse.

A person struggling with body dysmorphia may end up scrutinising their own jawline, skin texture, symmetry, waist, eye area or proportions against an image that has no coherent biological basis. Features can be pushed towards an ideal all at once, with none of the trade-offs that real bodies tend to involve. The result can look plausible at a glance while being impossible in literal terms.

That is a vicious reference point to put in front of anyone, especially someone whose brain is already primed to scan for defects. The comparison does not usually end with “that image is fake, so it means nothing”. It often lands closer to “why does my face not look right when this one does?” The sense of being uniquely wrong can deepen fast when the target has quietly stepped outside human reality.

The message underneath the image is brutally simple. If the body is treated properly, pushed hard enough and managed well enough, perfection should follow. AI makes that fantasy look cleaner and more believable because it removes every sign that a real body is a living system rather than a design project. That is dangerous territory. The problem is not only that the image is unrealistic. The problem is that it quietly supports a worldview in which the body should always submit.

Human models still carry context, even in a highly manufactured advert. There is at least some implied sense of genetics, routine, access, editing, surgery, professional support, money, time and circumstance. However artificial the final image may be, there is still a person somewhere behind the surface.

AI strips that away. There is no person to place in context, no lived body moving through time, and no human mess to help the mind make sense of what it is seeing. The image arrives as pure appearance, detached from any lived conditions that might explain it. That can make the comparison feel colder and more absolute. There is nothing to humanise the image or soften the standard. It sits there like a finished answer.

This is why the “it’s just advertising” defence is rubbish. Advertising does not merely sell products. It helps set background expectations about what bodies should look like, what counts as polished, what is desirable, and which kinds of appearance get quietly coded as failure. AI-generated models push that process further because they remove the last meaningful constraint on idealisation.

Once synthetic bodies become normal in marketing, there is no obvious stopping point. Each new image can be smoother, more symmetrical, more refined and more impossible than the last. Human models cannot compete with that. Human viewers cannot either.

AI-generated models do not simply continue the old beauty game with new software. They change the scale of it. The brain is still being asked to compare, judge and measure, while the standard it is measuring against no longer has to obey the basic terms of being human. That puts more pressure on everyone, and it puts the sharpest pressure on the people already carrying the greatest risk.

And that is the bit the industry keeps trying to glide past. The appeal of AI-generated models is not hard to understand. They are cheaper, easier to control and wonderfully compliant. They let brands build the perfect body without the inconvenience of a person. From a commercial point of view, that is efficient. From a mental health point of view, it is bleak.

Index:MH and the Body Image Safe Certification

Body Image Safe draws a clear line here. AI-generated models and heavily synthetic bodies don’t belong in marketing that claims to be responsible, because they train comparison against something unliveable. If a body does not exist, it should not be a beauty standard. Our standards push brands back toward reality: real people, real variation, and visuals that don’t rely on engineering “perfection” to sell.

Agree that AI-generated models don’t belong in fashion advertising? Add your name to back Body Image Safe standards and help us show the industry what shoppers want.