Skip to content

Research report · October 2026 · 20 min read

Create with people. Not instead of them.AI in fashion campaigns: how far it has gone, who is losing the work, what is being lost, the rules, and what we propose

Brands and fashion houses are already replacing the photographers, models, stylists and crews behind their campaigns with generated images, and the tools they use are sold on exactly that promise: the shoot, without the people. What comes out is an image made only by AI, with no trace of anyone who would have made it and no name on it. This report sets out how far that has gone, how many people it is costing work, and what is lost when a campaign is typed instead of made. It closes with what The Campaign Studios proposes instead: AI as the production tool, with the work of real creative people inside the studio, every one of them credited by name and recorded in a Campaign Passport.

Published 25 October 2026 · By The Campaign Studios · 7 sections · 65 sources · 9 figures

Abstract

What this report finds, in one paragraph.

Fashion and retail brands, from the high street to the luxury houses, have moved from experimenting with AI-generated images to running AI campaigns: in American Vogue, in collections shown on Vogue Runway, and in the marketing of at least nine luxury and designer houses. This report brings together 60 outside sources from 2023 to 2026: executive surveys, brand statements, the websites of AI fashion-image tools, member surveys of photographers and other creative workers, consumer and academic research, and court and regulatory decisions. It finds that the tools are sold, in their own words, as the shoot without the shoot, and that what they make carries no trace of the people who would have made it: none of the eight we reviewed says it credits them by name. The cost falls on those people first. 58% of the photographers answering one professional body's survey report lost commissions, a quarter of illustrators in another report lost work, and one projection puts €22 billion of creators' revenue at risk over five years; for makeup artists, hairstylists and stylists there are statements, but no counts yet. What is lost is not only work but the making itself: in US law a prompt is not authorship, and in one experiment work written with AI ideas grew more alike. Brands pay too, in trust. The report closes with what The Campaign Studios proposes in answer: AI as the production tool, with the creative work coming from people. Every contributor is credited by name in a Campaign Passport, which records who made what and what it is licensed for; a real person is cast only with their written consent; and nothing is made from a prompt.

Key findings

In the order of the sections. Findings 01 to 11 are taken from numbered sources; 12 and 13 are what the studio proposes and does today. Surveys are of people who chose to answer; projections are not measurements.

  • 2 · How far brands have goneMore than 35% of fashion executives already use generative AI in areas including image creation, and Zalando says around 70% of its editorial campaign images in late 2024 were AI-generated. [4] [5]
  • 2 · How far brands have goneGuess's AI advertisement in the August 2025 American Vogue carried one production line, "Produced by Seraphinne Vallora on AI", and no other reported credit. [9] [15]
  • 2 · How far brands have goneAt least nine luxury and designer houses have used AI-generated images, film or models in their marketing since 2023, among them Valentino, Gucci, Prada, Moncler and Burberry; David Koma showed two collections on generated images. [18] [20] [34] [33] [36] [21]
  • 2 · How far brands have goneThe AI fashion-image tools sell the shoot without the people: Botika offers "Without a single shoot", and FASHN images made "instead of a studio, photographer, and models". Of the eight tools we reviewed, none says it credits the people behind the look by name. [39] [40]
  • 3 · Who loses the workAOP members reporting lost commissioned work rose from 30% (September 2024) to 58.1% (early 2025), and was 58% in January 2026, with an average reported loss of about £34,900. [27] [47]
  • 3 · Who loses the workA quarter of illustrators report lost work; only 10% of over 5,500 UK creative workers expect AI to help their industry; 99.6% of Equity's screen members voting are prepared to refuse digital scanning; CISAC projects €22 billion of creators' revenue at risk over five years. [48] [49] [19] [50]
  • 3 · Who loses the workMakeup artists and hairstylists are named by the Model Alliance among the fashion workers raising concerns about AI; no survey yet counts their lost work. [8] [28]
  • 3 · What is lostA prompt is not authorship: in the US, "the mere provision of prompts" does not make AI output copyrightable. In one experiment, stories written with AI ideas were more similar to each other than stories by people alone. [13] [51]
  • 4 · What it costs brandsAlmost 90% of consumers want to know whether an image was made with AI; 62% of UK consumers say they trust marketing less since brands started using AI. [52] [56]
  • 4 · What it costs brandsDisclosing that an ad was made by AI lowered purchase intention in one experiment, and made luxury brands look less authentic in another, because AI ads seemed to take less effort. [54] [55]
  • 5 · The rulesNew York requires separate written consent for a model's digital replica; EU disclosure duties apply from 2 August 2026. [14] [23]
  • 6 · Our proposalWe propose AI as the production tool and creative people as the authors: whoever's work is chosen is credited by name in the campaign's Campaign Passport, which records who made what, where every part came from and what it is licensed for. [59] [60]
  • 6 · How it worksDesigners, makeup artists, hairstylists, stylists, photographers and set designers send their work into the studio, and a person reviews every piece; a real person is cast only after agreeing in writing; and there is no prompt: a campaign is directed visually. [59] [61] [62]

Figure 1

Four numbers from the sources

70%

Around this share of Zalando's editorial campaign images were AI-generated in the fourth quarter of 2024. [5]

58%

AOP members reporting lost commissioned work to generative AI, January 2026. [47]

€22bn

Creators' revenue projected at risk from generative AI over five years, music and audiovisual. [50]

2

Names reported on the Guess advertisement in Vogue: the brand and the agency's production line. [15]

Figure 1. Headline numbers, each in its source's own terms: a retailer's own figure (Zalando, reported by Reuters, May 2025), a member survey (AOP, January 2026), a projection, not a measurement (CISAC, December 2024) and the credits reported on one advertisement (August 2025). Sections 2 and 3 give the detail.Sources: [5] [47] [50] [15]

1 · Introduction

A campaign image is made by a team

A fashion campaign has never been the work of one person. Behind a single photograph there is usually a photographer and their assistants, a stylist, a makeup artist, a hairstylist, a set designer, a casting director, the models, a retoucher, and the designer whose piece is being shown. Most of them are named in the credits when the work is published, because the credit is how the next job is won.

Generative AI changes what that photograph costs, and who is in it. In 2023 McKinsey estimated that generative AI could add between $150 billion and $275 billion to the operating profits of the apparel, fashion and luxury sectors within three to five years.1 The same year, Goldman Sachs economists estimated that generative AI could expose the equivalent of 300 million full-time jobs worldwide to automation.2 That is a measure of exposure, not a count of jobs lost, and the same economists pointed out that new technology has historically created new work as well as removing it.2 The question this report asks is where the gains are going, and what happens to the people who used to make the image.

It asks it in four parts. How far have brands and fashion houses gone (section 2)? Who is losing the work, and what is lost with it (section 3)? What does it cost the brands themselves, in trust and in rights (section 4)? And what are the rules now asking for (section 5)? Section 6 then sets out what The Campaign Studios proposes in answer: what it is, how it works today, and why it matters to us. The methodology note at the end explains how the sources were chosen and where the evidence is thin.

2 · How far brands have gone

From experiment to the pages of Vogue

From the high street to the luxury houses: adoption in numbers, the brands that have done it, and what reached the page.

2.1 Adoption in numbers

No source we found counts how many brands now make campaign images with AI. The closest regular measure is the executive survey behind The State of Fashion, published each year by The Business of Fashion and McKinsey & Company. In the 2024 edition, 73% of fashion executives said generative AI would be a priority for their businesses that year, but only 28% had tried it in creative processes for design and product development.3 In the 2026 edition, more than 35 percent of executives reported already using it in areas such as online customer service, image creation, copywriting, consumer search or product discovery.4 The survey does not separate campaign imagery from the other uses.

Figure 2

Fashion's adoption of generative AI

Panel A: executive surveys. Panel B: one retailer's own figure.

A. Fashion executives, The State of Fashion surveys

Priority for 202473%
Tried it in design28%
Already using it, 2026more than 35%

B. Zalando: editorial campaign images that were AI-generated, Q4 2024

Zalandoaround 70%
Figure 2. Panel A: in the 2024 edition, 73% of executives said generative AI would be a priority in 2024 and 28% had tried it in creative processes for design and product development; in the 2026 edition, more than 35% reported already using it in areas such as customer service, image creation, copywriting, search or product discovery. The bases differ: these are not a single series, and none of them is a count of brands using AI for campaign images. Panel B: Zalando told Reuters that generative AI also cut image production from six to eight weeks to three to four days, and costs by 90%.Sources: [3] [4] [5]

2.2 The brands that have done it

In March 2023, Levi Strauss & Co. announced that it would experiment with AI-generated models alongside its human models. Critics asked why the brand would not simply book more real models; the company answered that the AI models would supplement its shoots with live models, not replace them, and denied that saving money was the motive.6 In July 2024 Mango released its first AI-generated campaign, for its teen line, and said publicly that AI had been used.7 In March 2025 H&M told CNN that it planned to create 30 "digital twins" of its models that year, with their consent, and that the models would own the rights to their twins and be paid each time one was used.8

At Zalando, Europe's large online fashion retailer, the change is already measured in volume. In May 2025 the company told Reuters that around 70 per cent of its editorial campaign images in the fourth quarter of 2024 had been generated with AI, and that generative AI had cut image production from six to eight weeks to three to four days, and its cost by 90 per cent.5 In August 2025 Guess ran advertisements whose models were generated by an AI agency in the August issue of American Vogue.9 And in December 2025 City AM reported that Zara was using AI to edit images of models from past campaigns into new clothes and new locations, after asking those models for permission; two of them said they had been offered the same fee as a standard shoot. A spokesperson for the retailer said it uses AI to complement its existing processes.10

Figure 3

From experiment to campaign, and the rules that followed

Brand and fashion-house moments, and the responses of workers, courts and lawmakers, 2023 to 2026.

202320242025202612345678910111213141516171819
BrandsWorkers and the law
  1. 122 March 2023BrandsLevi Strauss & Co. announces it will experiment with AI-generated models alongside its human models. [6]
  2. 2July 2024BrandsMango releases its first AI-generated campaign, for its teen line, and says AI was used. [7]
  3. 312 July 2024Workers and the lawThe EU AI Act is published in the Official Journal. Article 50 sets transparency duties for AI-generated content. [11]
  4. 416 November 2024BrandsCoca-Cola's AI-generated remake of its 1995 Christmas advertisement draws a backlash online. [12]
  5. 529 January 2025Workers and the lawThe US Copyright Office: AI outputs are protected only where a human author determined sufficient expressive elements; prompts alone are not enough. [13]
  6. 6March 2025BrandsH&M says it plans 30 digital twins of its models that year, with their consent, owned by the models and paid for each use. [8]
  7. 77 May 2025BrandsZalando tells Reuters that around 70% of its editorial campaign images in the fourth quarter of 2024 were AI-generated. [5]
  8. 819 June 2025Workers and the lawNew York's Fashion Workers Act is in force: separate written consent before a model's digital replica is created or used. [14]
  9. 9August 2025BrandsGuess runs advertisements with AI-generated models in American Vogue; the fine print reads “Produced by Seraphinne Vallora on AI”. [9] [15]
  10. 10August 2025BrandsJ.Crew's campaign images are revealed to be AI-generated; a digital artist is credited only after the backlash grows. [16]
  11. 114 November 2025Workers and the lawThe High Court in London rules mostly for Stability AI in Getty Images' claim; permission to appeal is later granted. [17]
  12. 122 December 2025BrandsValentino posts an AI-generated video for its DeVain handbag, labelled as made using AI; the BBC reports widespread criticism in the comments. [18]
  13. 13December 2025Workers and the lawEquity's indicative ballot: 99.6% prepared to refuse digital scanning on set, on a 75.1% turnout. [19]
  14. 1418 December 2025BrandsCity AM reports that Zara uses AI to dress images of models from past campaigns, with their permission. [10]
  15. 1525 February 2026BrandsGucci posts AI-generated images labelled “created with AI” before Demna's first show for the house, and is criticised online. [20]
  16. 1625 June 2026BrandsDavid Koma's Resort 2027 lookbook, shown on Vogue Runway, uses an AI-generated model. [21] [22]
  17. 172 August 2026Workers and the lawAI Act Article 50 transparency duties apply. [23]
  18. 189 October 2026BrandsDavid Koma presents Spring 2027 as an image series made with photography and generative AI. [24]
  19. 192 December 2026Workers and the lawMachine-readable marking applies to generative systems already on the market. [23]
Figure 3. Dated events named in this report. Where a source gives only a month, only the month is shown; on the strip those events sit near mid-month. Items 17 and 19 are dates set in law that apply from the day shown.Sources: [6] [7] [11] [12] [13] [8] [5] [14] [9] [15] [16] [17] [19] [10] [18] [20] [21] [22] [24] [23]

2.3 The Vogue page: one production line

The Guess advertisement is the clearest picture of what a generated campaign looks like when it reaches print. The two-page spread showed models who do not exist, made by Seraphinne Vallora, a London-based AI marketing agency.9 The only sign that the people in the pictures were not real was a small note in the corner of the page,25 which read "Produced by Seraphinne Vallora on AI".15 Vogue said the advertisement came from its advertising department, not its editorial team.26 The agency's founders said they value photography, stylists and everyone involved in a photo shoot, and that their work is meant to supplement, not replace.9

Set against a conventional campaign, the difference is in the credits. A campaign shot with people names the people who made it. The coverage of the Guess page reports two names: the brand, and the agency's production line. None of the reports we read mentions a credited photographer, stylist, makeup artist, hairstylist, set designer or model.

Figure 4

Credits compared

The people the sources name around a fashion shoot, and the credit printed on the Guess advertisement in Vogue.

Around a shoot, as named in the sources

  1. Photographer[27] [28]
  2. Stylist[9] [29]
  3. Makeup artist[8] [28]
  4. Hairstylist[8] [28]
  5. Model[27] [28]
  6. Retoucher[27]
  7. Lighting assistant[27]
  8. Studio[27]

On the Guess page in Vogue, August 2025

  1. Guessthe brand
  2. “Produced by Seraphinne Vallora on AI”[15]

No other credit is reported in the coverage we read.

Figure 4. Left: roles named by the AOP's chief executive, the Model Alliance, a makeup artist and the agency's own founders when describing who works on a shoot. Right: the credit on the advertisement as reported. The agency's founders said they still value photography, stylists and everyone involved in a shoot; the coverage does not report who, if anyone, worked on the images beyond the agency.Sources: [8] [9] [15] [27] [28] [29]

2.4 Not only the high street: the fashion houses

Most of the brands above sell to the mass market. Luxury and designer houses have used generated images too, and have met the same reaction. In December 2025 Valentino posted an AI-generated video on Instagram for its DeVain handbag, part of what it called a "digital creative project" with digital artists. The video carried a label saying it was made using AI. The BBC reported that many of the hundreds of comments under it criticised the use of AI, and a trends analyst told the BBC that audiences often read AI in a brand's visual identity as "cost-saving disguised as innovation".18

In February 2026, days before Demna's first runway show for Gucci in Milan, the house posted AI-generated images on social media, clearly labelled "created with AI". The BBC reported that critics called them "AI slop" and asked why a high-fashion company needed cost-cutting technology for its marketing, while some users praised them.20 Asked by CNN whether his use of AI was controversial, Demna answered: "I don't think so! I think this is 2026. I'm using things as a tool."30

The London designer David Koma has used generated images to present whole collections. For his Resort 2027 lookbook, Vogue Runway reported in June 2026 that Koma "created an AI visualization of his initial, imagined muse" to show the looks.21 WWD wrote that the model in the lookbook "is not human but generated using new digital tools", and quoted the designer: "Digital tech is another creative medium within my process."22 His Spring 2027 collection, published in October 2026, was presented as an image series "created through a dialogue between photography and generative AI", according to the house's collection notes as reported by Tom + Lorenzo.24 Vogue Runway's reviews of both collections name no photographer, stylist, hair or makeup artist, or model.2131 We found no press coverage of how audiences responded. Comments under the label's own Instagram posts announcing the collection include criticism of the AI model (observed by The Campaign Studios, October 2026).

They are not alone. Moncler's 2023 campaign with adidas Originals, The Art of Explorers, paired "AI-generated adventurers" with mixed-media sculptures,32 and in 2025 the house was the brand in "an experimental spot created with Veo", an AI film the agency R/GA made for Google in four weeks.33 Prada's fragrance campaign of April 2023 fed photographs of its bottles, taken by the still-life photographer Johann Besse, into an AI image generator, and kept the AI's errors as alternative versions of the final images.34 In 2026 the house "admitted that artificial intelligence (AI) was deployed in the campaign's post-production" of its Spring/Summer 2026 film, which had been shot with a 50-person crew.35 In its review of the year, the trade publication Glossy reported that Jil Sander "incorporated AI-generated visuals into a broader conceptual campaign" in February 2025, that MCM used AI imagery "as extensions of traditional campaigns", and that Burberry experimented with "AI-enhanced movement" on imagery from a 1980s campaign. Hugo Boss told Glossy it has used AI imagery across its e-commerce platforms since 2023, and in January 2025 launched AI-powered content on its global e-commerce platform.3637

Counting only cases reported by a named publication, at least nine luxury and designer houses have used AI-generated images, film or models in their marketing since 2023: Valentino, Gucci, Prada, Moncler, Burberry, Jil Sander, MCM, Hugo Boss and David Koma. The count is a floor, not a census. Several of these houses used AI alongside photographers, artists and crews, as Prada and Moncler did, rather than in place of them. That difference matters, and section 6 returns to it.

2.5 The tools: sold as the shoot without the people

Behind many of these images is a fast-growing group of AI fashion-image tools, sold directly to brands, and the way they describe themselves says what they are for. On 24 October 2026 we read the websites of eight of them: Botika, FASHN, Photoroom, Modelia, Claid, Lalaland.ai, Caimera and Pixelcut. Lalaland.ai, an AI-model service, is now part of Browzwear, a company that "specialises in 3D fashion design software" and acquired it in July 2025;38 the address lalaland.ai now opens Browzwear's page for those AI models.

Most lead with the same promise. Botika's homepage reads "Shoot your collection. Without a single shoot", and "From flat lay to campaign, without a shoot".39 FASHN offers "A full fashion photoshoot, without the studio", describes a virtual photoshoot as images made "with AI instead of a studio, photographer, and models", and invites brands to "Direct every scene without a crew". Asked on the same page whether AI can replace a traditional fashion photoshoot, it answers: "For most catalog and everyday content, yes."40 Photoroom shows clothes on AI-generated models with "no photo shoot required", and names what it takes out of the way: "Coordinating models, photographers, and stylists slows every product drop".41 Modelia describes its generator as "No photoshoot or studio needed".42 Claid tells brands to "Forget about managing and paying for traditional production".43 The page lalaland.ai now opens offers "lifelike, validated AI models", and lists among their benefits "No photoshoots, no buyout or license fees".44

What these pages describe is specific. From a product photograph, the tool makes an image in which the person is generated, and with the person go the hair, the makeup, the styling, the set and the light. The image is only AI. Nobody who would have made those choices is in it, and nothing in it says who they would have been.

Some take care where real people are concerned. FASHN says its terms do not allow misleading images of real people made without consent.45 Caimera offers a cast drawn "from a pool of AI or Real fashion models".46 But none of the eight says, in the pages we read, that it credits by name the people whose work is in the image: the photographer, the stylist, the makeup artist, the hairstylist, the set designer, the designer or the model.

We believe that is the wrong aim. The question is not whether to use AI, but whether it leaves the people out of the work or brings them in. We want it to bring them in, because their choices are what make the work real; section 6 sets out how.

3 · Who loses the work

The people behind the image lose the work first

Photographers, models, makeup artists, hairstylists, stylists, crews and other creators: what the surveys count, what they do not, and what is lost beyond the work.

Every generated image in section 2 is a shoot that did not happen, and a shoot is a day's paid work for a room full of people. The numbers below come from the people's own professional bodies and unions. They are not complete, and they do not all ask the same question, but they point the same way: more than half of the commercial photographers answering one survey have lost commissions, a quarter of illustrators in another have lost work, only one in ten of the creative workers in a third expects AI to be good for their industry, and a global projection puts €22 billion of creators' income at risk over five years.

3.1 Photographers

The clearest numbers come from the people closest to the camera. The Association of Photographers (AOP), the UK's professional body for commercial photographers, has surveyed its members repeatedly. In September 2024, 30 per cent said they had lost commissioned work to generative AI. By early 2025 the figure was 58.1 per cent.27 Its chief executive, Isabelle Doran, told Creative Review that clients were choosing AI to cut costs, and that most photographers were not losing their careers outright so much as watching commissions shrink until paid work became very hard to find. The effect, in the same interview, reaches the people around a shoot too: models, retouchers, lighting assistants and studios.27 In the AOP's January 2026 survey, answered by around 600 members, 58 per cent again reported lost commissioned work, and the average reported loss per person rose from about £14,400 the previous year to about £34,900.47

Figure 5

Photographers: lost commissions, and what they cost

Association of Photographers member surveys, UK.

A. Share of members reporting lost commissioned work to generative AI

September 202430%
Early 202558.1%
January 202658%

B. Average reported loss per person

Previous yearabout £14,400
January 2026about £34,900
Figure 5. Panel A: 30% in September 2024 and 58.1% in early 2025, as reported by Creative Review; 58% in the January 2026 survey of around 600 members. Panel B: the January 2026 survey's average reported loss per person, against the previous year's. Members who chose to answer; not a sample of all photographers.Sources: [27] [47]

3.2 Models

When Levi's announced its plan in 2023, Sara Ziff, founder of the Model Alliance, warned NBC News that fit models could be replaced by AI body scans.6 In a preliminary Model Alliance poll of over 100 fashion models and influencers, an overwhelming majority expected AI to harm their careers, and about one in five said they had already been asked to undergo body scans.28

3.3 Makeup artists, hairstylists and stylists

We found no survey that measures lost work for makeup artists, hairstylists or fashion stylists on their own; the evidence for them is statements, not numbers. The Model Alliance says that models and other fashion workers, including makeup artists, hair stylists and photographers, have raised concerns about losing work and about their images being altered with AI without consent or pay.28 When H&M announced its digital twins, Ziff said the initiative had the potential to replace a host of fashion workers, and named makeup artists and hairstylists among them.8 A makeup artist who works on brand and editorial shoots told Sourcing Journal in June 2025 that losing that work would be "truly heartbreaking", and that it was hard to support a company that uses AI in place of a creative team.29 When the model is generated, the people who would have done the hair, the makeup and the styling are usually removed with the model.

3.4 Crews, illustrators, performers and other creators

Crews. In a Bectu survey of over 5,500 creative workers in the UK, published in June 2025, just 10% predicted that AI would have a positive impact on their industry over the next five years.48

Illustrators. In a Society of Authors survey of 787 members published in April 2024, a quarter of illustrators said they had already lost work to generative AI.49

Performers. In December 2025, UK actors' union Equity announced the result of an indicative ballot of its film and television members: 99.6 per cent, on a turnout of 75.1 per cent, said they were prepared to refuse to be digitally scanned on set until AI protections were secured. The union is asking for protections built on explicit consent, clear terms and fair pay for use.19

Music and screen. A study commissioned by CISAC, the international confederation of authors' and composers' societies, and published in December 2024, projected that music creators could see 24 per cent and audiovisual creators 21 per cent of their revenues at risk by 2028: a cumulative loss of €22 billion over five years if the rules stay as they are.50

Figure 6

Creators' revenues at risk by 2028

Projected share of revenues at risk from generative AI, CISAC global study.

Music creators24%
Audiovisual creators21%
Figure 6. A projection, not a measurement: a cumulative loss of €22 billion over five years if the rules stay as they are. Study commissioned by CISAC and carried out by PMP Strategy, published December 2024.Source: [50]

3.5 What is lost when the image is typed

The work is not the only thing that goes. A campaign used to be made by people choosing: the photographer chose the light, the stylist the silhouette, the makeup artist the skin, the set designer the room. With a prompt-based tool, those choices become a sentence typed into a box, and the tool makes the rest.

The law has already drawn a line here. The US Copyright Office concluded in January 2025 that AI output can be protected only where a human author has determined sufficient expressive elements, which can include creative arrangements or modifications of the output, "but not the mere provision of prompts".13 In other words, typing the description does not make the typist the author of the picture. Research points the same way for the work itself. In an experiment published in Science Advances in July 2024, short stories written with ideas from an AI model were rated more creative and better written, but they were "more similar to each other than stories by humans alone": in the authors' words, "an increase in individual creativity at the risk of losing collective novelty".51 The study is about writing, not photography, and it does not show that prompting ends creativity. It does show what is at stake when everyone draws from the same machine.

It is our view, and we state it as a view, that prompting has taken a great deal of creativity out of image-making. People no longer make the picture. They describe it, and the machine decides the rest. A campaign made that way can look finished and still have nobody's eye in it. That is why the studio we propose in section 6 has no prompt at all.

4 · What it costs brands

Cheaper to make, harder to trust

4.1 What consumers say

The saving is real; so is the doubt. In research published by Getty Images in April 2024, almost 90 per cent of consumers worldwide said they wanted to know whether an image had been created using AI, 98 per cent agreed that authentic images and videos are pivotal in establishing trust, and 76 per cent said they were getting to the point where they could not tell whether an image was real.52 In December 2024, NielsenIQ reported a study of more than 2,000 people, about 150 of them measured by EEG: viewers identified most AI-generated ads, found them less engaging and more "annoying", "boring" and "confusing" than traditional ads, and even ads judged high quality produced weaker memory activation in the brain.53

A 2025 peer-reviewed experiment with 358 consumers found that telling people an advertisement had been created by AI lowered both the value they saw in it and their intention to buy, because it made the advertisement less credible.54 A study of luxury brands in the Journal of Advertising Research found the same direction: when a brand disclosed that it used AI, consumers viewed it less favourably, because AI advertisements were believed to "require less effort, making them feel inauthentic".55 For a house whose price rests on craft, that is a direct cost. In September 2026, 62 per cent of 1,000 UK consumers surveyed for Optimizely said they trusted marketing content less since brands started using AI, and 67 per cent said they found AI-generated images of people in marketing and advertising creepy.56

Figure 7

What consumers say about AI images in marketing

Share of respondents, three surveys.

A. Getty Images, global, published April 2024

Can't tell if an image is real76%

B. Optimizely, 1,000 UK consumers, September 2026

AI images of people are creepy67%
Trust marketing less62%

C. Clutch, 601 consumers, August 2026

Trust less if AI misleads64%
Less likely to buy53%
Figure 7. Each panel is a separate survey with its own question and sample; the bars are not comparable across panels. The same Getty Images release reports that almost 90% of consumers want to know whether an image was created using AI, and Clutch that nearly 90% say AI content should be disclosed. Stated attitudes, not measured purchases.Sources: [52] [56] [57]

4.2 When it becomes public

Brands have learned this in public. In November 2024, Coca-Cola's AI-generated remake of its 1995 "Holidays Are Coming" advertisement was called "creepy", "dystopian" and "soulless" by viewers online.12 In August 2025, J.Crew's campaign images were revealed to be AI-generated; the brand had not disclosed it, critics called the pictures "AI slop", and a digital artist was credited only after the backlash had grown.16 After the Guess advertisement ran in Vogue, ABC News reported that some readers were cancelling their subscriptions and calling for a boycott.58 No count of cancellations has been published.

The stakes are not only reputational. In a survey of 601 consumers in August 2026, 53 per cent said they were less likely to buy from brands they know use AI in social media content, and 64 per cent said they would trust a brand less if its products were misrepresented with AI-generated content.57 These are stated intentions. None of the sources we found shows a brand's sales falling because of AI imagery, and we do not claim it. What they show is lower stated trust, and in one experiment, lower intention to buy once an advertisement is disclosed as AI-made.

4.3 One image, no credits, uncertain rights

A generated campaign also leaves gaps on paper. As section 3.5 set out, the US Copyright Office has concluded that the mere provision of prompts is not enough for copyright.13 A brand that generates its campaign from prompts alone may therefore own less of it than it assumes. The law on the material these systems learn from is also unsettled: in November 2025 the High Court in London rejected most of Getty Images' claims against Stability AI, finding for Getty only on a narrow trade mark point, and permission to appeal has since been granted.17

Put together, the evidence describes a particular kind of campaign: one image, made quickly and cheaply, disclosed in small print or not at all, with no names on it and uncertain rights behind it. Consumers say they want to know how an image was made. A licensed, credited campaign can tell them, because it records who made it and on what terms.

5 · The rules are changing

Consent, disclosure and authorship are becoming law

Consent. New York's Fashion Workers Act, in force since 19 June 2025, requires model management companies to obtain clear written consent, separate from the representation agreement, before a model's digital replica is created or used, setting out its scope, purpose, rate of pay and duration. It also prohibits altering a model's digital replica with AI without the model's separate written consent.14

Disclosure. In the European Union, Article 50 of the AI Act sets transparency duties for AI-generated content. Those who publish AI-generated or manipulated images that resemble real people or events (deep fakes) must disclose that they were artificially generated; where the work is evidently artistic or creative, the duty is limited to disclosing that such content exists, in a way that does not spoil the work. The providers of generative systems must also mark their outputs in a machine-readable way.11 These duties apply from 2 August 2026, with a later date, 2 December 2026, for the machine-readable marking of generative systems that were already on the market.23

Authorship. In the United States, copyright follows the human author, not the prompt (sections 3.5 and 4.3).13

These rules do not ban AI in campaigns. They set a floor: consent before a person's likeness is used, disclosure of how an image was made, and authorship that rests with people. We think that floor should be the starting point of how an AI production tool is designed, not an afterthought added to it. Section 6 sets out our proposal: a studio built on that floor.

6 · Our proposal

AI production that brings creative people in

Sections 2 to 5 set out the problem: campaigns made without the people who would have made them, at a cost to those people and to the brands, under rules that now ask for consent, disclosure and a human author. This section sets out what The Campaign Studios proposes instead, how it works today, and why it matters to us.

6.1 What we propose

We propose a different use of the same technology. In The Campaign Studios, AI is the production tool, and the creative work comes from people: designers, makeup artists, hairstylists, stylists, photographers and set designers, whose work is chosen, credited and recorded in every campaign it appears in. AI production is becoming extraordinarily powerful, and too much of the conversation around it asks how many people can be removed from the process. We start from the opposite question: how much more can creative people make, and be seen making, when technology removes the barriers around them? Four commitments follow from it, and each is built into the studio today (section 6.2).

People make the work. The pieces, the makeup and hair looks, the styling and the sets in a campaign can come from real creative people, and a person reviews every piece that is sent in. AI makes the photograph; it does not replace the team.

Every contributor is credited by name. Whoever's work is chosen is named in the campaign's Campaign Passport, in the role they played, and the credit stays with the campaign.

Consent comes first. A real person is cast only after agreeing in writing, on rules they set themselves, and anything they have not answered is treated as no.

Nothing is made from a prompt. A campaign is directed visually, the way a shoot is, so the choices in it stay with the people making it.

6.2 How it works in The Campaign Studios

Everything below is what the studio does today, as its own published pages describe it.

Real creative work, in every room. The Contributor Network is open to the people section 3 counts. Designers send pieces: fashion, accessories, jewellery, shoes and headwear. Makeup artists send makeup looks, designed in the studio's Beauty Designer. Hairstylists send hair looks. Stylists send Styled looks, complete outfits built in the studio's Dressing and Styling rooms. Photographers and set designers send Worlds: a photograph of an empty room, which the studio finishes into a set. A person looks at every piece, and not all of it is accepted.59

Every contributor, credited by name. When a brand chooses a contributor's work, the contributor's name is written into that campaign's Campaign Passport, in the role they played.59 The Campaign Passport is the record kept with each campaign: what it was made from, who is credited, and how its photographs were made. It shows who made what, where every part came from and what it is licensed for, and each credit is kept as it stood on the day the campaign was frozen.60

Real people, cast with their consent and on their own rules. Through Casting, real people can apply to become a Studio Model: a digital representation built from their photographs, at the studio's cost. Nothing is made from their photographs unless they are accepted and have agreed in writing, one statement at a time, after seeing the actual pictures. They set five permissions, from makeup and hair to swimwear, lingerie and alcohol, each starting at No, and the studio never takes on gambling, tobacco or political work for anyone. They can withdraw at any time, and every campaign they appear in records them in its Campaign Passport, under their cast name.61

No prompt window. There is nothing to type. A campaign is directed the way a shoot is: the cast, the look, the pose, the camera and the light are each chosen visually. In the studio's own words: "Creative people do not think in prompt syntax. They think in casting, styling, light and camera."62

Free to join. Making a profile, sending work and applying through Casting are free. In the founding period, taking part is for the credit and the work: contributors are not charged and are not paid.5961

The Contributor Network is new, and we say so. Makeup looks and Worlds are reviewed now but are not yet open to every brand.59 We will say more only when it is real.

6.3 The difference, side by side

Put an image from a generation tool next to a campaign made in The Campaign Studios, and the difference is not in how the picture looks. It is in what stands behind it. The first follows what we call the replacement model: AI makes the image instead of the people who would have made it. The second follows the collaboration model: AI is the production tool, and the creative work comes from people who are named. Brands can choose either. H&M's stated approach, consent and payment for each use of a model's twin, is closer to the second, and Equity welcomed the pledge to pay while asking for protections in union agreements and in law.8 The technology does not choose. People do.

Figure 8

The difference, side by side

What stands behind an image from an AI generation tool, and behind a campaign made with The Campaign Studios.

An image from a generation tool

  1. How it is madeFrom a prompt or a photo
  2. The people in itNo trace of anyone
  3. CreditsNone, or one line
  4. Where it came fromNot shown
  5. ConsentLeft to the tool's terms
  6. RightsUncertain

A campaign made with The Campaign Studios

  1. How it is madeDirected, not prompted
  2. The people in itEach contributor traced
  3. CreditsEach, by name and role
  4. Where it came fromEvery part recorded
  5. ConsentIn writing, before anything
  6. RightsEach part licensed

Where something was not recorded, the Campaign Passport says not recorded.

Figure 8. Left: the campaigns and tool pages reviewed in sections 2.3 to 2.5, not every use of AI; some tools ask for consent where a real person is used, but none we reviewed says it credits the people behind the look. "Uncertain": in the US, AI output is copyrightable only where a human author determined enough of it, and prompts alone are not enough. Right: as the studio's own pages describe it; in a fully digital campaign, the studio's own cast, makeup, hair and Worlds are credited to Campaign Studio (section 6.6). The law points to the right-hand column: separate written consent for a model's digital replica in New York, disclosure in the EU, and copyright that rests with a human author in the US (sections 3.5 and 5).Sources: [15] [39] [40] [45] [13] [14] [11] [60] [59] [61]

6.4 Credit for the people the industry forgets

Credit matters most to the people who have the least of it. Even in physical shoots, emerging designers often lend their pieces in the hope of exposure. The independent fashion platform 1 Granary describes designers' red-carpet looks as "typically loaned (read: worn for free)", with designers fronting "all of the shipping and alteration costs" and paid, instead, in exposure.63 The exposure does not always arrive with a name on it. In 2023 the designer Tarun Tahiliani said a television series had presented garments his studio provided in good faith as the work of "a fictitious designer (actor) representing a fictitious label", and called it "a shocking breach of faith".64 If it can happen to an established name, it happens more easily to a new one. It happened to our founder: pieces lent to shoots and published without the designer's name, because the label was not yet big enough to be named.

With the Campaign Passport, the credit is always there. It is saved with the campaign, in black and white: who made it, and who contributed. A designer's piece is credited to the designer, a makeup look or hair look to its artist, a Styled look to its stylist, a World to the photographer or set designer who sent it.60 A designer does not even send a sample: the studio works from photographs of the piece.59 Nobody is hidden. And if a campaign is great, everyone in it gains the exposure and the credit, not only the brand that published it.

Figure 9

What a Campaign Passport records

The same record for a campaign made with real talent and for a fully digital one.

Who is credited

  1. Creative directionYour brand
  2. The pieceBrand or designer
  3. CastBy cast name
  4. MakeupArtist or studio
  5. HairArtist or studio
  6. StylingStylist or brand
  7. WorldIts maker or studio

How it was made

  1. ProductionAI-assisted
  2. AI model usedBy name
  3. PhotographsContact sheet
  4. WhenThe day it was made
  5. LicenceFor your campaigns

Where something was not recorded, it says not recorded.

Figure 9. In a real Campaign Passport, names stand where this figure shows roles. "Studio" means Campaign Studio, credited for the studio's own cast, makeup, hair and Worlds. Each credit is copied in when the work is chosen and kept as it was when the campaign was frozen. The Campaign Passport records who contributed; it does not replace the licence, which the studio's Terms of Service set out in full.Sources: [62] [60]

6.5 Why it matters to us

The Campaign Studios was founded by a fashion designer, a graduate of London College of Fashion, University of the Arts London. It was built from inside the industry this report describes: the photographers, models, stylists, makeup artists, hairstylists, set designers and young designers who build their careers one credit at a time.

In that industry the credit is the currency. A name under a photograph is how a stylist is found by the next brand, how a makeup artist's book fills, how an emerging designer is first noticed. An image with no names on it does not only take away a day's fee. It takes away the next job too.

Our founder has lent pieces to shoots and seen them published without the designer's name (section 6.4). That is why credit is not a feature added to the studio but the ground it is built on, and why the studio cannot work like the tools in section 2.5. We do not believe the future of creative work should have fewer creative people in it.

6.6 If you work alone

A brand does not have to collaborate. The studio's own libraries hold more than 2,500 assets across the Cast, Worlds, campaigns, garments, styling, makeup, hair and poses, included with the subscription and licensed for use in the brand's campaigns.6260 Its own curated campaigns can be taken as they are: choose one, and its cast, beauty, styling and World come with it, and the brand brings the collection. The studio's Cast are AI-made characters, not real people.65

Even then, there is no prompt. We believe prompting has taken creativity away from image-making, so the studio does not ask for one: you direct, visually, as you would on set. And the Campaign Passport still records where every part came from and what it is licensed for. The studio's own cast, makeup, hair and Worlds are credited to Campaign Studio, the brand is credited for its own piece and creative direction, and the AI model the photographs were made with is named, because being clear about the method is part of being clear about the people. Where something was not recorded, the Campaign Passport says not recorded rather than guessing.60

A campaign photograph made in the studio: a model in a black sequinned dress against a concrete wall
The Concrete Wall, one of the studio's own curated campaigns, made in The Campaign Studios with AI. Taken as it is, its cast, beauty and World are the studio's own, and its Campaign Passport credits them to Campaign Studio.

7 · Conclusion

The technology does not choose. People do.

The evidence in this report points one way. Brands have moved from experiments to campaigns in the pages of Vogue, at least nine luxury and designer houses have followed, and one large retailer reported that around 70 per cent of its editorial campaign images in late 2024 were generated. The tools behind such images are sold, in their own words, as the shoot without the shoot, and what they make carries no trace of the people who would have made it. The cost falls first on those people: 58 per cent of the photographers answering the AOP's survey have lost commissions, illustrators, crews and performers report the same, and models, makeup artists, hairstylists and stylists say so in statements, even where no one has yet counted their losses. With the work goes the making: a prompt is not authorship, and work drawn from the same machine grows more alike. The brands, in turn, face consumers who say they trust AI imagery less, and laws that ask for consent, disclosure and a human author.

None of that is written into the technology. It is a set of choices about who is credited, who consents and who is left out. AI production can mean more people, from more places, making more work, with their names on it.

That is the choice The Campaign Studios proposes, and it is why the studio exists. AI is the production tool; people are the authors. The work of real creative people is inside the studio, every contributor is credited by name, a real person is cast only with their written consent, and there is no prompt to write: a campaign is directed, not described. And every campaign carries a Campaign Passport, so that whoever made it, and whatever it was made from, is never hidden.

Create with people. Not instead of them.

Methodology note

How sources were chosen. We used primary sources where we could reach them: the US Copyright Office, the EU's Official Journal, the published releases of Getty Images, NielsenIQ, Optimizely, Clutch, CISAC, Bectu and the Model Alliance, and peer-reviewed journal articles (Journal of Research in Interactive Marketing, Science Advances, Journal of Advertising Research). Where a primary source was not reachable, we used established news and trade publications (BBC News, NBC News, CNN, ABC News, Reuters, City AM, Vogue Runway, WWD, Adweek, Creative Review, Glossy, Just Style, FashionNetwork, Decrypt, Cosmetics Business, Retail Systems, The Edge, Outlook India, 1 Granary, Tom + Lorenzo) and law-firm summaries (Morgan Lewis, Baker McKenzie, Loeb & Loeb, CMS). Every number is given in its source's own terms, with its date.

Limits. Member surveys are answered by those who choose to answer. Projections and estimates of exposure are not measurements. Several consumer surveys were run or paid for by companies with an interest in the answer (an image library, a measurement firm, a software company and a ratings platform), and they ask different questions of different samples, so their figures are not comparable. Consumer figures are stated attitudes, not observed purchases; we found no source that measures a brand's lost sales from AI imagery. We found no survey that counts lost work for makeup artists, hairstylists or stylists separately, and none that counts uncredited designers. The Science Advances experiment is about short stories, not images. Some studies report that disclosing AI can be neutral or even raise trust in some settings; we could not verify them fully and have not cited them. McKinsey's site could not be opened on the day the links were checked; the State of Fashion figures were confirmed through its published page summary and matching trade coverage.

The fashion houses. The count of nine in section 2.4 includes only houses whose use of AI-generated images, film or models was reported by a named publication. For Jil Sander, MCM and Burberry the only report we found is Glossy's review of the year; for Prada's Spring/Summer 2026 film, The Edge's report of the house's own admission. We searched for other houses named in the press in connection with AI, and left out those where we found only unofficial, fan-made images or no report of the brand's own use. Reactions on social media are reported only where a news organisation described them, and no individual comment is quoted. The one exception is David Koma, for whom we found no press coverage of audience reaction: the sentence on comments under the label's Instagram posts is a first-hand observation by The Campaign Studios in October 2026, with screenshots held on file. We do not quote or name any commenter, and we give no count.

The tools. For section 2.5 we read the public websites of eight AI fashion-image tools on 24 October 2026: Botika, FASHN, Photoroom, Modelia, Claid, Lalaland.ai, Caimera and Pixelcut. Browzwear makes 3D garment design software; its acquisition of Lalaland.ai, a maker of AI-generated fashion models, was reported in July 2025, and the address lalaland.ai now opens Browzwear's page for those AI models, which is the page we quote. We read each homepage and, where it had one, its main page for AI photoshoots or AI models. We did not read every page, their full terms or their help centres, and we did not use the tools. Quotations are copied from the pages as they stood that day and may since have changed. "None says it credits" means we found no such statement in the pages we read, not that no such feature exists. Pixelcut was reviewed but is not quoted: its homepage makes no comparable claim about shoots. We also looked for VModel, Veesual, ZMO, Onbrand and CREATE studio. VModel's site now offers a general AI-model API and Veesual's an image-animation product, so neither was counted; ZMO's site could not be opened on the day because its security certificate had expired; and we could not identify the companies called Onbrand and CREATE studio with confidence.

Views and experience. Two statements in this report are ours rather than a source's, and are marked as such: that prompting has taken creativity out of image-making (sections 3.5 and 6.6), which is the studio's view; and the account of pieces published without the designer's name (section 6.4), which is our founder's own experience.

The studio. Section 6 describes The Campaign Studios as its own published pages describe it on 25 and 26 October 2026.6259616560 The photograph in section 6 is one of the studio's own campaigns. The report uses no stock photographs.

References

Numbered in order of first citation in the text. All links checked 24 and 25 October 2026. Dates are the publication dates shown by each source.

  1. 1"Generative AI: Unlocking the future of fashion", McKinsey & Company, 2023. $150 billion to $275 billion added to apparel, fashion and luxury operating profits in three to five years. www.mckinsey.com ↩
  2. 2"Generative AI could raise global GDP by 7%", Goldman Sachs Insights, 5 April 2023 (from the report by Joseph Briggs and Devesh Kodnani). The equivalent of 300 million full-time jobs exposed to automation. www.goldmansachs.com ↩
  3. 3Laura Husband, "Signal: Generative AI worth risk to amplify fashion's creative process", Just Style, 14 December 2023, reporting The State of Fashion 2024 (The Business of Fashion and McKinsey & Company). 73% of fashion executives said generative AI would be a priority in 2024; 28% had tried it in creative processes for design and product development. www.just-style.com ↩
  4. 4"The State of Fashion 2026: When the rules change", The Business of Fashion and McKinsey & Company, November 2025 (month from matching coverage, CXOToday, 18 November 2025). More than 35 percent of executives already use generative AI in areas such as online customer service, image creation, copywriting, consumer search or product discovery. www.mckinsey.com ↩
  5. 5Helen Reid, "Zalando uses AI to speed up marketing campaigns, cut costs", Reuters, republished by Internet Retailing, 7 May 2025. Around 70% of editorial campaign images AI-generated in the fourth quarter of 2024; production time from six to eight weeks to three to four days; costs down 90%. internetretailing.com.au ↩
  6. 6Sara Ruberg, "Backlash against AI supermodels triggers wider fears in fashion workforce", NBC News, 30 April 2023. Levi's plan with an AI model company announced 22 March 2023; Levi's statement that the AI models would supplement, not replace, live shoots; Sara Ziff (Model Alliance) on fit models. www.nbcnews.com ↩
  7. 7"Mango debuts first AI-generated campaign", FashionUnited, July 2024 (the page shows no date; July 2024 is from its address and matching coverage). Mango's first AI-generated campaign, for its teen line, and its disclosure of AI use. fashionunited.com ↩
  8. 8Issy Ronald, "Fashion giant H&M plans to use AI clones of its human models. Not everyone is happy", CNN, republished by ABC7, 1 April 2025. H&M's 30 digital twins, consent, ownership and pay; Sara Ziff naming makeup artists and hairstylists; Equity general secretary Paul W. Fleming's response. abc7.com ↩
  9. 9Kelly McCarthy, "Controversy stirs over AI generated 'models' in new Guess ads appearing in Vogue", ABC News, republished by ABC7, 1 August 2025. Guess advertisements with AI-generated models in the August 2025 issue of American Vogue, made by a London-based AI marketing agency; the agency founders' statements. abc7.com ↩
  10. 10Saskia Koopman, "Zara turns to AI-edited models amid shop closures", City AM, 18 December 2025. Zara uses AI to edit images of models for its e-commerce; models from past campaigns asked for permission; two models said they were offered a standard shoot fee; the retailer's spokesperson on complementing existing processes. cityam.com ↩
  11. 11Regulation (EU) 2024/1689 (the AI Act), Article 50, "Transparency obligations for providers and deployers of certain AI systems", Official Journal of the EU, 12 July 2024. eur-lex.europa.eu (also: artificialintelligenceact.eu) ↩
  12. 12Pranav Dixit, "'Death of art': Coca-Cola faces backlash over AI-generated Christmas ad", Business Today, 16 November 2024. A remake of the 1995 "Holidays Are Coming" advertisement with AI-generated visuals; viewers calling it "creepy", "dystopian" and "soulless". businesstoday.in ↩
  13. 13"Copyright Office releases Part 2 of Artificial Intelligence report", US Copyright Office, NewsNet issue 1060, 29 January 2025. Outputs of generative AI protected only where a human author has determined sufficient expressive elements; the mere provision of prompts is not enough. copyright.gov ↩
  14. 14"New York State enacts Fashion Workers Act", Morgan Lewis, 8 January 2025. In force 19 June 2025; separate written consent for a model's digital replica, with scope, purpose, rate of pay and duration; no AI alteration of a digital replica without separate written consent. www.morganlewis.com ↩
  15. 15Natasha Jokic, "Here's what's happening with AI models inside the newest issue of Vogue", BuzzFeed, 25 July 2025. The fine print on the advertisement: "Produced by Seraphinne Vallora on AI." www.buzzfeed.com ↩
  16. 16Brian L. Heidelberger, "Two recent GenAI controversies highlight brand/PR risks of use", Loeb & Loeb, 29 August 2025. J.Crew's campaign images revealed as AI-generated, no initial disclosure, a digital artist credited after the backlash grew. www.loeb.com ↩
  17. 17"Getty Images v Stability AI", CMS artificial intelligence and copyright case tracker (United Kingdom). Judgment of 4 November 2025: secondary copyright infringement claim rejected; trade mark claims succeeded only on a narrow point; permission to appeal granted 17 December 2025. cms.law ↩
  18. 18Liv McMahon, "Valentino criticised over 'disturbing' AI handbag ads", BBC News, 2 December 2025. An AI-generated Instagram video for the DeVain handbag, labelled as made using AI, part of a "digital creative project" with digital artists; many of the hundreds of comments criticising the use of AI; Anne-Liese Prem (Loop) and Dr Rebecca Swift (Getty Images) on the reaction. www.bbc.co.uk ↩
  19. 19"UK performers prepared to take industrial action over AI in landslide vote", C21Media, 18 December 2025. Equity indicative ballot: 99.6% yes, 75.1% turnout; protections based on consent, transparency of terms and fair pay. www.c21media.net ↩
  20. 20Liv McMahon, "Gucci criticised for 'AI slop' images ahead of major fashion show", BBC News, 25 February 2026. AI-generated images promoting the Milan show, clearly labelled "created with AI"; critics on social media, some users in favour; the BBC contacted Gucci and Kering for comment. www.bbc.co.uk ↩
  21. 21Joan Kennedy, "David Koma Resort 2027 Collection", Vogue Runway, 25 June 2026. For the lookbook, Koma "created an AI visualization of his initial, imagined muse" to show the looks in front of a blank backdrop. The review names no photographer, stylist, hair or makeup artist, or model. www.vogue.com ↩
  22. 22Tianwei Zhang, "David Koma Resort 2027", WWD, 25 June 2026. "AI-generated imagery"; the model in the lookbook "is not human but generated using new digital tools"; Koma: "Digital tech is another creative medium within my process." wwd.com ↩
  23. 23"New EU guidance on AI transparency: what should companies be doing from 2 August 2026", Baker McKenzie (Connect On Tech), 2 July 2026. Article 50 applies from 2 August 2026; under the AI Omnibus agreement, generative systems already on the market have until 2 December 2026 for machine-readable marking; draft Commission guidelines. connectontech.bakermckenzie.com ↩
  24. 24"David Koma Spring 2027 Collection", Tom + Lorenzo, 9 October 2026. The collection presented as an image series "created through a dialogue between photography and generative AI"; images courtesy of David Koma. tomandlorenzo.com ↩
  25. 25"AI-generated models appear for first time in 'Vogue' magazine", Capture, 30 July 2025. The small note in the corner of the advertisement as the only indication of AI. www.capturemag.com.au ↩
  26. 26Samuel Gut, "Vogue US faces backlash over Guess ad featuring AI-generated model", FashionNetwork, 1 August 2025. Vogue's statement that the advertisement came from its advertising department, not its editorial team. us.fashionnetwork.com ↩
  27. 27Laura Havlin, "How AI is threatening commercial photographers", Creative Review, 14 August 2025. AOP surveys: 30% (September 2024) to 58.1% (early 2025) of members reporting lost work; Isabelle Doran on clients, commissions and the wider crew. www.creativereview.co.uk ↩
  28. 28"Research", Model Alliance (page undated; its key-findings summary is dated September 2025). Models and other fashion workers, including makeup artists, hair stylists and photographers, raising concerns about AI; a preliminary poll of over 100 models and influencers; about one in five already asked to undergo body scans. www.modelalliance.org ↩
  29. 29Meghan Hall, "Are models and creatives facing digital displacement?", Sourcing Journal, republished by Yahoo Tech, 18 June 2025. A makeup artist on brand and editorial work and on brands that use AI in place of creative teams; Sara Ziff on job replacement beyond models. tech.yahoo.com ↩
  30. 30Rachel Tashjian, "Can Gucci make luxury AI?", CNN, 27 February 2026. The AI-generated images on Gucci's Instagram, a tag added to say they were generated with AI; Demna: "I don't think so! I think this is 2026. I'm using things as a tool." edition.cnn.com ↩
  31. 31Anna Cafolla, "David Koma Spring 2027 Ready-to-Wear Collection", Vogue Runway, 9 October 2026. The review of the collection; it names no photographer, stylist, hair or makeup artist, or model, and does not discuss how the images were made. www.vogue.com ↩
  32. 32Stephanie Hirschmiller, "Adidas and Moncler collab features AI 'adventurers' and NFTs", Decrypt, 2 October 2023. The campaign "The Art of Explorers" pairs "AI-generated adventurers" with mixed-media sculptures; earlier teaser images with "AI-generated models". decrypt.co ↩
  33. 33Brittaney Kiefer, "How R/GA used AI to make an 'impossible' ad", Adweek, 8 August 2025. "An experimental spot created with Veo and featuring luxury fashion brand Moncler", made by R/GA for Google in four weeks and unveiled in June. www.adweek.com ↩
  34. 34Austyn King, "Prada taps into AI for latest fragrance campaign", Cosmetics Business, 18 April 2023. Bottles photographed by still-life photographer Johann Besse, then fed into an AI image generator; "The social media campaign also features the AI's errors as alternative versions of the final imagery." cosmeticsbusiness.com ↩
  35. 35Kong Wai Yeng, "Should the fashion industry finally embrace AI imagery?", Options, The Edge Singapore, 6 May 2026 (first published in The Edge Malaysia, 13 April 2026). Prada "admitted that artificial intelligence (AI) was deployed in the campaign's post-production" of its Spring/Summer 2026 film, shot with a 50-person crew. www.optionstheedge.com ↩
  36. 36Zofia Zwieglinska, "In 2025, luxury fashion's AI marketing experiments hit a turning point", Glossy, 24 December 2025. Jil Sander "incorporated AI-generated visuals into a broader conceptual campaign" (February 2025); MCM "used AI imagery to create stylized, surreal environments as extensions of traditional campaigns" (spring 2025); Burberry "continued to experiment with AI-enhanced movement, leveraging imagery from a 1980s campaign"; Hugo Boss says "it has been using AI imagery across e-commerce platforms since 2023". www.glossy.co ↩
  37. 37Silvia Iacovcich, "Hugo Boss 'puts AI into action' with image and video creation", Retail Systems, 24 January 2025. "Hugo Boss has launched artificial intelligence-powered content on its global e-commerce platform". retail-systems.com ↩
  38. 38Jangoulun Singsit, "Browzwear acquires AI fashion model firm Lalaland", Just Style, 28 July 2025. Browzwear, based in Singapore, "specialises in 3D fashion design software"; Lalaland.ai, founded in Amsterdam in 2019, specialises in AI-generated fashion models and "offers a cost-effective alternative to traditional photoshoots". www.just-style.com ↩
  39. 39Botika, homepage (botika.io now opens botika.com), accessed 24 October 2026. Headlines "Shoot your collection. Without a single shoot" and "From flat lay to campaign, without a shoot". botika.com ↩
  40. 40"AI Fashion Photoshoot for Clothing Brands", FASHN, accessed 24 October 2026. "A full fashion photoshoot, without the studio"; a virtual photoshoot "creates on-model fashion images with AI instead of a studio, photographer, and models"; "Direct every scene without a crew"; to "Can AI replace a traditional fashion photoshoot?", "For most catalog and everyday content, yes." fashn.ai ↩
  41. 41"AI-Generated Fashion Models For Product Photography", Photoroom, accessed 24 October 2026. Products shown on AI-generated models, "no photo shoot required"; "Coordinating models, photographers, and stylists slows every product drop". www.photoroom.com ↩
  42. 42Modelia, homepage, accessed 24 October 2026. Page description: "No photoshoot or studio needed". modelia.ai ↩
  43. 43"AI fashion photoshoot", Claid.ai, accessed 24 October 2026. "Forget about managing and paying for traditional production"; "campaign-ready on-model images without a traditional shoot". claid.ai ↩
  44. 44"Browzwear Sell", Browzwear, accessed 24 October 2026: the page the address lalaland.ai now opens, offering "a curated collection of lifelike, validated AI models". Among their benefits: "No photoshoots, no buyout or license fees." browzwear.com ↩
  45. 45FASHN, homepage, accessed 24 October 2026. Frequently asked questions: images of a real person need their permission, and FASHN's terms do not allow misleading images of real people made without consent; FASHN does not add a visible watermark or label. fashn.ai ↩
  46. 46Caimera, homepage, accessed 24 October 2026. "AI Visual Production Platform for Fashion Teams"; "Create Your Own Cast from a pool of AI or Real fashion models". www.caimera.ai ↩
  47. 47"Survey shows that 58% of photographers have lost commissioned work to generative AI", Videomaker, February 2026 (the page shows no date; month from its image path), reporting the AOP's January 2026 member survey, first reported by PetaPixel on 30 January 2026. About 600 respondents; average reported loss about £34,900, up from about £14,400. www.videomaker.com (also: petapixel.com) ↩
  48. 48"Bectu endorses report calling for protections for creative workers from the impacts of generative AI", Bectu, 4 June 2025. Bectu's survey of over 5,500 creative workers: 10% predicted AI will have a positive impact on their industry over the next five years. bectu.org.uk ↩
  49. 49Melina Spanoudi, "A third of translators report losing work to generative AI systems, SoA survey reveals", The Bookseller, 12 April 2024. Society of Authors survey of 787 members; a quarter of illustrators lost work to generative AI. thebookseller.com ↩
  50. 50"Global economic study shows human creators' future at risk from generative AI", CISAC press release (study by PMP Strategy), 4 December 2024. 24% and 21% of revenues at risk by 2028; €22 billion cumulative. www.cisac.org ↩
  51. 51Anil R. Doshi and Oliver P. Hauser, "Generative AI enhances individual creativity but reduces the collective diversity of novel content", Science Advances, vol. 10, no. 28, 12 July 2024, doi:10.1126/sciadv.adn5290. Stories written with AI ideas were judged more creative, but "are more similar to each other than stories by humans alone", "an increase in individual creativity at the risk of losing collective novelty". doi.org ↩
  52. 52"Nearly 90% of consumers want transparency on AI images finds Getty Images report", Getty Images press release, 30 April 2024 ("Building Trust in the Age of AI", VisualGPS research with MarketCast; surveys of 7,500 adults each across 25 countries). Almost 90% want to know whether an image was created using AI; 98% say authentic images and videos are pivotal to trust; 76% say they can no longer tell if an image is real. newsroom.gettyimages.com ↩
  53. 53"NIQ research uncovers hidden consumer attitudes toward AI-generated ads", NIQ (NielsenIQ) press release, 12 December 2024. More than 2,000 participants, about 150 measured by EEG; AI-generated ads seen as less engaging and more "annoying", "boring" and "confusing"; weaker memory activation even for high-quality ads. nielseniq.com ↩
  54. 54Hien Bui, "Examining the effect of AI advertising involvement disclosure on advertising value and purchase intentions", Journal of Research in Interactive Marketing, vol. 20, no. 5, 2025, pp. 613 to 632, doi:10.1108/JRIM-02-2025-0066. Experiment with 358 consumers: disclosing that an advertisement was created by AI lowered advertising value and purchase intentions, through lower credibility. ray.yorksj.ac.uk ↩
  55. 55Rita Ngoc To, Yi-Chia Wu, Parichehr Kianian and Zhe Zhang, "When AI doesn't sell Prada: why using AI-generated advertisements backfires for luxury brands", Journal of Advertising Research, vol. 65, no. 2, 2025, pp. 202 to 236, doi:10.1080/00218499.2025.2454120. Summarised by Tarleton State University, 17 March 2025: when AI use was disclosed, consumers viewed the brand less favourably, because AI ads were believed to "require less effort, making them feel inauthentic". doi.org (summary: www.tarleton.edu) ↩
  56. 56"Six in 10 consumers say AI has made brand marketing less trustworthy", Optimizely press release, London, 22 September 2026. Survey of 1,000 UK consumers and 100 marketers: 62% trust marketing content less since brands started using AI; 67% find AI-generated images of people in marketing and ads creepy; 72% more likely to engage with content that feels human. www.optimizely.com ↩
  57. 57"Clutch report: 53% of consumers are less likely to buy from brands using AI on social media", Clutch press release, 8 September 2026. Survey of 601 consumers, August 2026: 53% less likely to purchase; 64% would trust less if products were misrepresented with AI; nearly 90% say AI content should be disclosed. clutch.co ↩
  58. 58"AI model in ad sparks backlash at Vogue", Good Morning America (ABC News), 1 August 2025. The Guess advertisement prompting some readers to cancel subscriptions and call for a boycott. video.goodmorningamerica.com ↩
  59. 59"Become a Contributor", The Campaign Studios, accessed 25 October 2026. "Real work. Real credit."; "No shoot to book, no samples to send"; five kinds of work open (pieces, makeup looks, Worlds, hair looks, Styled looks); credit written into the campaign's Campaign Passport when a founder selects the work; every piece looked at by a person; "A profile is free. Contributors are not charged and are not paid."; makeup looks and Worlds not open to every founder. thecampaignstudios.com ↩
  60. 60"Campaign Passport", The Campaign Studios, accessed 25 October 2026. "A record kept with each campaign: what it was made from, who is credited, and how its photographs were made"; contributors credited in their role; the studio's own cast, hair, makeup and worlds credited to Campaign Studio; "not recorded" where nothing was recorded; "Licensed for your campaigns"; it does not replace the licence. thecampaignstudios.com ↩
  61. 61"Casting", The Campaign Studios, accessed 27 October 2026. Studio Models built from a person's photographs at the studio's cost; agreement in writing, one statement at a time; the actual pictures seen first; five permissions, each starting at No; gambling, tobacco and political work never taken on; withdrawal at any time; credit in the Campaign Passport under the cast name; "Nothing is charged and nobody is paid today." thecampaignstudios.com ↩
  62. 62"How it works", The Campaign Studios, accessed 26 October 2026. "Direct. Don't describe."; "Creative people do not think in prompt syntax"; "Custom or curated"; "Work with real creatives credited by name, or use more than 2,500 library assets across the Cast, Worlds, campaigns, garments, styling, makeup, hair and poses, included with your subscription"; the Campaign Passport's rows, ending "Licence: Licensed for your campaigns". thecampaignstudios.com ↩
  63. 63"It's time independent designers get fair compensation for red-carpet looks", 1 Granary, 30 June 2024. Looks "typically loaned (read: worn for free)"; designers "front all of the shipping and alteration costs"; payment "comes in the form of exposure". 1granary.substack.com ↩
  64. 64Garima Das, "Fashion designer Tarun Tahiliani slams 'Made in Heaven 2' makers for using his outfits without giving credit", Outlook India, 18 August 2023. Garments his studio provided in good faith presented on screen by "a fictitious designer (actor) representing a fictitious label"; "a shocking breach of faith". www.outlookindia.com ↩
  65. 65"The Libraries", The Campaign Studios, accessed 24 October 2026. The studio's campaigns and libraries; "The studio's Cast are AI-made characters, not real people"; take one campaign, with its cast, beauty, styling and world, or direct your own, "Not open to every account". thecampaignstudios.com ↩
Open StudioBecome a contributor