Projektdetaljer
Beskrivelse
AI-generated fashion models are rapidly entering commercial use across e-commerce, social media, and marketing campaigns. These range from fully synthetic models to digital replicas of real models, whose appearance can be manipulated and redistributed across campaigns (Mateescu, West, & Pinto, 2025). In many instances, artificial fashion modeling falls under the category of 'mimetic products': products that explicitly refer to a "real" counterpart while acknowledging that they are not real. As with products such as fake meat, fur, wood, and musical instruments, the mimetic product typically attempts to resemble key attributes of a real referent, yet its value also depends on differentiating itself from what it imitates. Recent research shows how this tension between mimicry and differentiation constitutes a fundamental strategic dilemma for producers of mimetic products, who may pursue pure imitation or augmenting strategies with divergent consequences for market development (Nelson, 2025).
Advances in generative AI are, however, rapidly closing the gap between the mimetic product and its real referent. It has become increasingly difficult to distinguish AI-generated fashion models from real human models, whereby the category boundary between "real" and "artificial" models is not only contested but also eroding. Because conventions defining product categories evolve in response to new entrants (Krabbe & Grodal, 2023), the very criteria by which audiences judge what counts as a "real" model are shifting. This is accelerated by the way producers of AI fashion models can, on demand and with increasing precision, replicate markers of humanness and authenticity. For instance, human models who lean into visible "realness" such as freckles, wrinkles, blemished skin, and stretch marks to differentiate themselves from their synthetic counterparts find that the very attributes they deploy to signal authenticity easily become targets for imitation, through which the basis of differentiation erodes. But while the technology erodes the boundary between real and artificial, regulation now push to legally reestablish the boundary through mandatory labeling of AI-generated content. E.g. the EU AI Act imposes transparency obligations requiring that AI-generated content be clearly labeled as artificial, by which producers of AI fashion models face the paradox of being mandated to disclose the very artificiality their technology is designed to render imperceptible. Consequently, such disclosure itself becomes a new site of positioning, contestation, and audience judgment.
This raises important questions that extant research has not yet addressed in detail. For instance, how does the category of artificial fashion models emerge and gain footing in markets when the imitation technology upon which it is based gradually erodes the boundary to its real referent? What are the consequences for product positioning, authenticity, and audience demands in the age of a social media platform economy? How do new regulatory demands for transparency and labeling shape how producers justify and position artificial fashion models? And how do the power dynamics between brands, agencies, and fashion models shape the terms on which this new category takes hold? In this project, we are guided by such questions.
The project builds upon a qualitative archival case study, in which we analyze advertisements on social media and in fashion magazines together with interviews with AI fashion model agencies, marketers, and fashion industry stakeholders. The project advances our understanding of how new mimetic product categories emerge when imitation technologies blur the boundary between the artificial and the real. In doing so, the project aims to contribute to research on category emergence, mimicry strategies in creative industries, and authenticity issues arising as mimetic products reshape cultural ideals of beauty and appearance in marketing.
References
Krabbe, A. D., & Grodal, S. (2023). The aesthetic evolution of product categories. Administrative Science Quarterly, 68(3), 734–780.
Mateescu, A., West, Z., & Pinto, S. (2025). Fashion’s data doubles: How AI is reshaping modeling work. ILR Worker Institute & Data & Society.
Nelson, A. J. (2025). “Even better than the real thing”?: Electronic organs and the dilemma of product mimicry. Strategic Management Journal, 1–25
Advances in generative AI are, however, rapidly closing the gap between the mimetic product and its real referent. It has become increasingly difficult to distinguish AI-generated fashion models from real human models, whereby the category boundary between "real" and "artificial" models is not only contested but also eroding. Because conventions defining product categories evolve in response to new entrants (Krabbe & Grodal, 2023), the very criteria by which audiences judge what counts as a "real" model are shifting. This is accelerated by the way producers of AI fashion models can, on demand and with increasing precision, replicate markers of humanness and authenticity. For instance, human models who lean into visible "realness" such as freckles, wrinkles, blemished skin, and stretch marks to differentiate themselves from their synthetic counterparts find that the very attributes they deploy to signal authenticity easily become targets for imitation, through which the basis of differentiation erodes. But while the technology erodes the boundary between real and artificial, regulation now push to legally reestablish the boundary through mandatory labeling of AI-generated content. E.g. the EU AI Act imposes transparency obligations requiring that AI-generated content be clearly labeled as artificial, by which producers of AI fashion models face the paradox of being mandated to disclose the very artificiality their technology is designed to render imperceptible. Consequently, such disclosure itself becomes a new site of positioning, contestation, and audience judgment.
This raises important questions that extant research has not yet addressed in detail. For instance, how does the category of artificial fashion models emerge and gain footing in markets when the imitation technology upon which it is based gradually erodes the boundary to its real referent? What are the consequences for product positioning, authenticity, and audience demands in the age of a social media platform economy? How do new regulatory demands for transparency and labeling shape how producers justify and position artificial fashion models? And how do the power dynamics between brands, agencies, and fashion models shape the terms on which this new category takes hold? In this project, we are guided by such questions.
The project builds upon a qualitative archival case study, in which we analyze advertisements on social media and in fashion magazines together with interviews with AI fashion model agencies, marketers, and fashion industry stakeholders. The project advances our understanding of how new mimetic product categories emerge when imitation technologies blur the boundary between the artificial and the real. In doing so, the project aims to contribute to research on category emergence, mimicry strategies in creative industries, and authenticity issues arising as mimetic products reshape cultural ideals of beauty and appearance in marketing.
References
Krabbe, A. D., & Grodal, S. (2023). The aesthetic evolution of product categories. Administrative Science Quarterly, 68(3), 734–780.
Mateescu, A., West, Z., & Pinto, S. (2025). Fashion’s data doubles: How AI is reshaping modeling work. ILR Worker Institute & Data & Society.
Nelson, A. J. (2025). “Even better than the real thing”?: Electronic organs and the dilemma of product mimicry. Strategic Management Journal, 1–25
| Akronym | AURA |
|---|---|
| Status | Igangværende |
| Effektiv start/slut dato | 02/03/26 → 29/02/28 |
Emneord
- Erhvervsliv, handel og økonomi
- Medier, kommunikation og sprog