AI Advertising Without Replacing Models or Creative Teams

I recently watched a short fashion film made with generative AI. It was excellent. Not excellent for AI. Excellent as visual work.

The concept, mood, movement, and edit showed real taste. It was better than anything I could currently produce with those tools.

Then I read the comments. One side called the creator a prompter who was replacing an entire production crew. The other called AI just another camera and suggested talented people had nothing to fear.

I do not think either position is honest enough.

I have spent years working as a fashion and editorial photographer for publications, sports teams, and modeling agencies. I know what a real set requires. I also know how often a strong idea dies before the first call sheet is written because the budget is not there.

Generative AI can help talented people create work that was previously out of reach. It can also reduce the amount of paid work available to other talented people. Both can be true.

AI is more than a camera, but it still needs a creative mind

A camera records a scene. Generative AI can help create the scene, location, wardrobe, lighting, set, movement, retouching, and visual effects. It can function more like a compact production studio than a single tool.

But dismissing the person behind strong AI work as a prompter is like calling a photographer a button pusher. The software does not choose the idea, emotional tone, composition, pacing, edit, or hundreds of rejected versions. Generative AI can produce attractive accidents. Consistently good work still requires judgment.

A better analogy is a power saw. The saw does not design the furniture or create the craftsperson’s taste. It does let one person work faster. That can make more ambitious work possible, but it can also reduce the number of labor hours a project requires.

The same tool can expand creativity and compress labor.

The concern about creative jobs is legitimate

If a brand had already budgeted for a photographer, model, stylist, makeup artist, assistants, studio, travel, and post production, then cancels that shoot and asks one person to generate the campaign, paid work has been removed.

Hiring a real model for likeness licensing does not automatically protect everyone else who might have worked on the set.

The International Labour Organization concluded in 2025 that job transformation is a more likely broad effect of generative AI than complete automation because most occupations still contain tasks requiring human input. Transformation can still mean fewer shoot days, smaller crews, different skills, and fewer entry level opportunities.^[5]

“The best people will always have work” is not a serious answer. Skilled professionals can lose assignments when clients change how they spend money.

Real campaigns already show several possible futures

H&M kept a real model and photographer in the process

In July 2025, H&M released campaign images featuring digital twins of real models. The company said the work was co-created with a dedicated team and released alongside a behind the scenes film. Photographer Johnny Kangasniemi described AI as another tool alongside his photography, while model Vanessa Moody emphasized the collaborative and transparent process.^[1]

The model did not disappear. The photographer did not disappear. AI extended what the team could make.

Zara used real models to create additional outfit imagery

Reuters reported in December 2025 that Zara was using AI to create new images of real models wearing different outfits. Zara’s parent company said the changes were mutually agreed with models and that compensation followed industry best practices. Reuters also cited an unnamed model who said they were paid the same amount they would have received for traveling to another shoot.^[2]

This is close to the model Spencer Models is building. A real person can approve the brand and usage, remain behind the identity, and receive compensation without traveling to every client or physically trying on every variation.

That does not end the labor debate. The Association of Photographers warned that fewer shoots could mean fewer commissions across the production ecosystem. The Zara example shows the opportunity and the risk in the same workflow.

Zalando shows why the efficiency is disruptive

Zalando told Reuters that generative AI reduced its imagery production time from roughly six to eight weeks to three or four days and reduced costs by 90 percent. It also said about 70 percent of its editorial campaign images in the fourth quarter of 2024 were AI generated.^[3]

Those are company reported figures, but they explain why the shift will not disappear. The question is whether the efficiency creates more campaigns, removes labor from existing campaigns, or does some of both.

Levi’s showed why messaging and trust matter

In 2023, Levi Strauss announced a partnership with Lalaland.ai to test AI generated models as a supplement to human models. After criticism, the company clarified that the pilot should not have been framed as a diversity solution and that it was not scaling back live shoots or its work with human collaborators.^[4]

Calling AI a supplement is not enough. Brands have to show who participates, what they approved, and how they are compensated.

Replacing a campaign is not the same as making one possible

Scenario	What happened	Likely effect
Replacement	A funded traditional shoot is canceled and recreated with a much smaller AI workflow	Existing assignments are displaced
Extension	A live shoot happens, then approved material is used to create more formats, locations, or campaign variations	Core work remains while the campaign produces more content
Market expansion	A smaller brand that could not afford a conventional campaign commissions a lower cost human backed production	Paid work may be created that otherwise would not exist

This is similar to what stock photography did for small businesses. It made professional imagery accessible to companies that could not afford custom production, while also replacing some custom assignments. AI goes further because it can make lower cost production feel customized.

The effect depends on which project is being compared with which alternative.

What a human first remote campaign could look like

Imagine a new jewelry brand in another state. It has a strong product and a limited launch budget. Flying a model and full crew to a location, building several sets, and shooting every product combination may be impossible.

A human first AI workflow could still involve:

A real model who approves the brand, creative scope, usage, and compensation
A photographer or capture team creating high quality source material
A product photographer or three dimensional artist documenting the actual jewelry
A creative director building the concept
An AI artist, retoucher, editor, or compositor producing the finished assets
The model, agency, and brand reviewing the work before release

That is not the same crew as a conventional location shoot. It is also not one anonymous prompt replacing all human involvement.

If the brand could never have afforded a traditional campaign, the smaller workflow may create paid work. If it canceled a fully funded shoot to use that workflow, it replaced work.

Same technology. Different economic reality.

Saving money does not lower the standard for truth

Remote production still has to represent the product honestly.

If AI changes the size of a pendant, the number of stones, or the metal finish, the ad may misrepresent the jewelry. If it makes clothing appear to drape, stretch, or fit in a way the real garment does not, the image can create a false impression.

The Federal Trade Commission says advertising claims must be truthful, cannot be deceptive or unfair, and should be supported by evidence.^[6]

A model illustrating a product is also different from a model claiming personal experience with it. A campaign should not make someone appear to say, “I wore this all day and loved the fit,” when that never happened. FTC guidance explains that a brand supplied image or message can be misleading when it communicates an experience the person did not have.^[7]

An AI label does not fix an inaccurate product or a fabricated endorsement.

What Spencer Models should stand for

Spencer Models is not being built as a synthetic face factory.

We want real models and creative professionals to remain part of the commercial relationship. AI should reduce unnecessary logistics, open campaigns to brands with smaller budgets, and extend the value of approved creative work.

Real models participate by choice. Traditional representation does not automatically authorize AI use.
Each campaign has defined rights. The brand, purpose, media, territory, term, compensation, and restrictions should be clear.
Creative work remains paid work. Capture, direction, product photography, editing, compositing, and quality control still have value.
The product must remain accurate. A beautiful image is not useful if it misrepresents what the customer will receive.
AI should not invent personal experience. A model illustrating a product is not automatically endorsing it.
Live production remains essential. Some campaigns need real movement, interaction, performance, fit, spontaneity, and the energy of people working together.

Not every campaign needs a plane ticket and a full production day. Not every campaign belongs inside a generator.

The question is not simply whether AI is good or bad

The useful questions are more specific:

Did the campaign replace funded human work, extend existing work, or create a project that otherwise could not happen?
Did the model understand and approve the use?
Who provided the creative direction and production work?
Who gets paid when the campaign runs?
Does the final asset accurately represent the product?
Is the audience being led to believe something that never happened?

AI will not preserve every creative role exactly as it exists today. It also does not have to reduce advertising to one person generating synthetic faces for brands that no longer hire anyone else.

The goal should not be to remove humans from advertising. It should be to remove barriers while keeping humans meaningfully involved in the value being created.

➤ Start a brand project

➤ Apply for representation

Frequently asked questions

Does AI advertising always replace creative jobs?

No. It can replace a funded production, extend an existing campaign, or make a smaller campaign possible when the traditional version was never affordable. The effect depends on the project and what would have happened without AI.

Can a real model appear in an AI campaign without physically wearing the product?

Yes, when the model has authorized the use and the campaign stays within the approved license. The product still must be represented accurately, and the ad should not imply that the model personally used or endorsed it when that is not true.

Is a digital twin the same as a synthetic model?

No. A digital twin or digital replica is based on a specific real person. A synthetic model is a generated identity that is not anchored to one represented human. Learn more in What Is an AI Augmented Model?

When should a brand still use a live shoot?

A live shoot is usually the better choice when the campaign depends on real performance, fit, movement, product interaction, interviews, spontaneous emotion, creator experience, or a physical event.

Written by Brian Spencer, founder of Spencer Models & Management, LLC. Brian’s background includes fashion and editorial photography for publications, national sports teams, and modeling agencies, as well as digital marketing, brand strategy, and AI assisted creative production. Last reviewed September 26, 2026.

Can AI Advertising Lower Costs Without Replacing Models and Creative Teams?