Over the past few years, hyper-realistic virtual influencers have transcended the boundaries between reality and fiction. Generated entirely by code, digital models have unlocked unprecedented opportunities for both fashion brands and the ecommerce stores which sell their wares. What happens when they operate entirely on artificial intelligence? 

To be clear, CGI has supported the fashion and modelling industries for years. Computer-generated imagery is a cost-effective and more sustainable approach to modelling clothes, reducing wasted cash and materials by giving brands design previews, lookbooks, and sample campaigns before committing to production.

AI images can deliver results within 5 minutes.

But artificial intelligence offers a new level of creative freedom for designers, giving them the power to create digital models at greater scale and speed. In March 2023, iconic denim brand Levi's announced a partnership with AI modelling agency Lalaland to use AI-generated models of different skin tones and body types. The move was met with both excitement and backlash, with critics arguing that the collaboration further marginalises underrepresented groups by excluding real people from the industry.

But whipping up AI mannequins for one-off campaigns is only the tip of the iceberg. The proliferation of AI modeling startups provides a glimpse of a future where any model wearing any clothing item can be generated with just a few prompts. Lalaland estimates that generating fashion images with AI images can deliver results within 5 minutes, with zero need for physical samples.

For small businesses and independent creators, this technology will serve as a leveling tool, empowering them to compete in markets traditionally dominated by resource-heavy rivals. Proprietors of virtual fashion shows also stand to benefit, not because their experience is necessarily higher quality than a real catwalk, but because they are much easier to deliver. Meanwhile, the collapsing barrier to design inspiring fashion is likely to give rise to a proliferation of new styles and approaches, catering to an increasing number of niches.

That said, it is unlikely that those resource-heavy rivals will be slow to insert artificial intelligence into their own workflows. Singaporean fashion photographer and virtual model creator Shavonne Wong expects that “the accessibility and affordability of AI tools will spark widespread adoption.”

However, she notes that some industries will still likely choose human models over their AI-generated counterparts. “I think it’s unlikely that AI models will completely replace human models, because certain campaigns, like skincare or body wash, are more suitable for humans,” she said, noting that, in addition to the healthcare dimension of those products, campaigns that involve physical interactions products or require less common types of poses may be too difficult for an AI model, pun intended, to deliver smoothly.

Beyond diversity in appearance, AI-generated faces may still create a pervasive uniformity. “There is a concern that AI models could lead to a homogenisation of beauty standards,” Shavonne says, “as AI-generated faces tend to conform to certain mathematical averages and may lack the unique quirks and imperfections that make human faces interesting.”

Not that models have a track record of creating relatable beauty standards, but AI risks overshadowing the subtle nuances that make each human face unique with a polished version that couldn’t exist in the real world. This shift could mold societal perceptions, fostering a constrained and stereotypical understanding of attractiveness.

‍

From a commercial standpoint, brands are focused on staying ahead of the curve, leaving them with an important question: how can they integrate AI-generated models into their workflows? 

Founder of digital fashion collective The Immersive Kind, Kadine James emphasises that there fashion brands can approach artificial intelligence in a range of ways. “By combining generative AI with virtual production and effects, we’ve used AI to create a catwalk video based only on a series of images, plus our archives of runway and front row videos,” she explains, noting that a first step for brands with AI could be as simple as using the tech to create digital lookbooks or to showcase products online.

“AI models could lead to a homogenisation of beauty standards.”

— Shavonne Wong, virtual model creator

AI models and their use in the fashion industry will be a marker of society’s relationship with beauty in a world where the lines between real and virtual are increasingly blurred. AI is not just another tool for the fashion industry's toolkit, it invites a seismic shift in how society perceives beauty, diversity, and even humanity itself.

It’s important to remember that this oncoming wave of creativity is no utopian vision: pixel-perfect pictures pose thorny questions from algorithmic bias to the erasure of human uniqueness. Brands navigating this terrain won't be just adopting a new technology, they'll be creating a new aesthetic language in which the choices they make today will affect not just commercial outcomes, but will echo in the societal norms of tomorrow.

“The accessibility and affordability of AI tools will spark widespread adoption.”

— Shavonne Wong, virtual model creator