Dressed in Data

In recent years, fashion designers have gained a multitude of options and use cases, with AI emerging as a powerful ally from the very first step: sketch creation.

Daniele Bardis, a London-based designer and fashion brand owner, uses generative AI to elevate his sketching process. He describes the images produced by Midjourney as of "stellar levels".

''You can easily go from zero to hero''

During my interview with Daniele, I sensed a deep interest towards AI. An interest he conveys clearly throughout his YouTube channel, showcasing how can fashion brands leverage AI tools to their advantage.

Daniele started 'Daniele Bardis' in 2013—a brand that specialises in crafting exceptional leather goods. The fashion designer moved from Rome to London in 2007, seeking greater opportunities. "I wanted to do things on a more international level," he told me in his Studio.

As we delved into the integration of AI in the creative process of his work, he spoke about the value it brought to his work. "When I discovered AI a year and a half ago, I immediately fell in love with it," he said. "I was obsessed with it because it was for me something new and unusual."

He shared his two favourite AI tools: Vizcom and Midjourney. While Vizcom is his go-to to refine sketches, Midjourney is a source of inspiration for photoshoots.

Daniele has a knack for online research, he likes to stay abreast of the latest trends. He discovered Vizcom and started using it to sketch his designs. "With a pencil and a white paper, on my train or in a coffee shop, I start doing my sketches then I run to my computer and insert it to Vizcom to get the virtual look of what I made."

Vizcom, founded in 2021, is an AI-powered creative tool originally designed for industrial designers. It can transform sketches of cars, bikes or bicycles into photo-realistic images. The tool is also used for fashion items like trainers, bags, jackets. "Even if you're not exceptionally good at sketching, you can photograph your sketch with a phone and send it to your computer. It will give you the virtual or 3D look," Daniele explained. "However, the better is your sketch, the better is your result."

The designer, who's an AI enthusiast got a bit concerned when he realised how accessible the tool is—even for children. One Saturday morning, he sat with his 8-year-old stepson and introduced him to Vizcom. Before he knew it, the child was able to generate good designs. "It's scary how good it can be, even with the prompts of a child," he recalls.

Despite his concern, he finds it now impossible to refrain from using the tool. The level of efficiency it brought to his workflow was beyond his expectations. "When I started developing the DNA of the bag: sketching, retracing in Illustrator, putting leather on the bags in Photoshop using online images, it was taking me forever. But with Vizcom, the entire process takes 10 min max," he explained.

In fashion design, perfectionism is a requirement. Attention to details and precision are hallmarks of good craftsmanship. Both designers are convinced that using AI can enhance human creativity. "Just don't do it too much," Gianluca cautioned.

"Fashion designers who use traditional methods will still fit in this world," Gianluca stated. "But the worry of job replacement is completely understandable." The conversation around designing with AI tools remains controversial. Danieles acknowledges that it will be difficult—somewhat impossible— to ditch the generative AI tools and revert to traditional methods. However, the two designers pointed out that humans and generative AI can work hand in hand to optimise their creative processes, under the condition of using it with the utmost caution; especially with the appearance of other promising tools, this time, described as AI-design assistants, the likes of AIDA.

AIDA—short for AI-based interactive design assistant—was developed in Hong Kong by Calvin Wong. Its creator explained in an interview for the South China Morning Post how the process works. "Designers upload to the tool the abstract images, the theme of the upcoming season, the fabric prints, the colours and the drawing sketches. Then designers just need to click design and the system can output each time eight outfits within 10 seconds." AIDA was officially launched with the FashionXAI show in Hong Kong and is currently available for designers in Europe and Asia Pacific.

Ultimately, It seems like with each passing day, a new innovation emerges, each more accessible and user-friendly than the last.

For other brands, generative AI is more than just an inspiration

As both small and big fashion brands began to showcase their collection for 2023/2024, it became evident that some of them were jumping on the AI-imagery trend to market their products.

This time, using a full AI-driven advertisement.

Fashion brands use generative AI in various aspects of their branding processes: Design and product development, Supply chain, Marketing, Digital commerce, Store operations, Organisation and customer experience.

The financial influence of generative AI in the industry is substantial. Over the next three to five years, generative AI could contribute an estimated $150 billion, and up to $275 billion to the apparel, fashion, and luxury sectors’ operating profits, according to a 2023 McKinsey analysis. 

However, the latest Business of Fashion (BOF) and McKinsey State of Fashion 2024 survey showcases that Marketing is the biggest use case for fashion executives.

Even though some fashion brands have used AI in different processes, those truly memorable—whether in a positive or negative light—are the ones that feature it as a centrepiece of their collection, in design or marketing.

The ones who are leveraging AI in their marketing strategy are engaging with their audience, receiving both negative and positive feedback: a debate found in the comments section of social media posts.

Casablanca, the French-Moroccan fashion house teamed up with photographer and AI artist Luke Nugent to introduce the Spring/Summer 2023 collection. Sheep.Inc, the London-based brand known for its interest in sustainable fashion, shared AI-generated on-model outfits on its official Instagram page—showing its keen interest in the experiment. Etro, the Italian fashion house, used generative AI to launch a dreamy fashion campaign for their 2024 spring collection. The campaign also marked the introduction of their first eyewear collection.

With the advent of Midjourney V6 or other tools used by brands' creative teams, it became possible to see a high level of photorealism associated with major fashion brands label.

After a closer inspection, one can possibly realise that it's AI. Yet, it's fair to say that, given the real-world scenarios emerging from each photograph, it's remarkably realistic for a virtual creation.

These campaigns and others demonstrate the creative potential of virtual photography.

Generative AI is entering the scene of fashion e-commerce

Anyone enjoying online shopping would come to accept that it's highly possible to miss the correct size. Yet, it's comforting to know that the return option is always available —or at least 90% of the time.

To optimise the customer experience and reduce the return rate, brands adopted AR technology. Simply put, AR or Augmented reality is explained by experts as a technology made for overlying digital images into the real world to enhance them with digital details.

It was in 1968 when the first AR technology was developed by computer scientist Ivan Sutherland. He created an AR head-mounted display system, laying the foundation for the use of AR technology.

So in the 2010s and early 2020s, retailers introduced the idea of visualising a garment on customers before buying it, with Converse being the first retailer to use the Virtual Try-on technology in 2012.

Luxury brands did embrace the trend. Burberry, Channel, and Prada have provided the virtual try-on option for a big selection of products, including eyewear, jewellery, and accessories.

Chorno24 a marketplace for watches, collaborated with famous watchmakers to enable watch enthusiasts to try their favourite watches virtually.

But, it's also possible to join AR technology with generative AI to take the virtual try-on innovation to the next level.

In fact, clothes are not static compared to jewellery and accessories, it's more complicated to fit the texture and the garment movement to the estimated body size. Still, some companies, like wanna fashion showcased the possibility to try on clothes virtually. They collaborated with famous brands like Diesel, Dolce&Gabanna and Valentino.

Matthew, who's been working on projects involving virtual try-ons, says there are different routes to do it. "Effectively, the concept involves scanning the body and then applying a virtual simulation of a garment on top of that. So looking at several anchor points on your body to begin to attach that digital garment to and generative AI can also play a role," he explained.

However, even with the new developments, he believes this technique can be well implemented to jewellery and footwear since the end points are fixed, but it's still challenging for body types.

"In fact, it still has a way to go before it gets to a point where people feel confident about it," he added.

At the heart of the new experiments happening in the sector, lies a factor that Matthew deems crucial: "emotion". For anyone engaging with these innovations or encountering them unexpectedly, be it at work or with their favourite brand; it's clear that they evoke something. It can be surprise, shock, confusion, or in my case: amusement.

Emotions lie at the centre of generative AI experiments

It's almost impossible to tackle the topic of AI without considering the emotions it evokes. The technology has the power to spark curiosity, excitement, and fascination. In other cases, scepticism and fear.

In 2023, when an image of the pope wearing a long white puffer Balenciaga-style jacket surfaced online, it gained widespread attention. At first glance, one can tell that it's an AI image. A deduction I probably made from having intentionally viewed dozens of such images lately. However, for those misled by the hyper-realistic nature of the image, the revelation that it was an AI creation came as a shock.

When the image went online, internet users had already seen Midjourney's potential. A week before the surfacing of the Pope's image, X was flooded with 50 hyper-realistic fake images of Donald Trump being arrested by the police of New York. Yet, the Balenciaga Pope jacket stole the spotlight during that period, due to its comedic character, according to a Forbes article.

However, generative AI is not always appreciated by the public. In many cases, it's perceived by the public as a threat to job security for models.

Levi's, the denim jeans powerhouse, was in the middle of an online backlash after the company announced its partnership with the AI-powered digital model studio: Lalaland.

According to Levi's press release, the initiative was introduced as a way to increase "the number of diverse models", which was heavily criticised by customers and online users, disapproving the use of virtual models instead of real diverse models. Levi's and Lalaland were aware of the backlash, and each one of them responded. Levi's updated their original statement to give more insights to the public and Lalaland has also clarified its position through interviews.

Models are at the forefront of the Technology X Fashion era

The fashion landscape has always been associated with models, as a symbol of elegance, beauty and confidence.

Throughout history, iconic figures graced runways in fashion capitals in the likes of Milan, Paris and New York; representing brands and showcasing the craftsmanship behind every design.

Today, AI-generated models are also gaining some sort of attention, spearheading the forefront of AI X fashion innovations.

While we discuss AI designs and AI-driven campaigns, the final look of the design or the advertising campaign is carried out by virtual models, which are crafted to showcase garments and fashion items in a digital sense.

Daniele Bardis shared that Midjourney puts designs on beautiful bodies and the garment is perfectly fitted on them. Something we can clearly witness from the images generated with AI designs, like the ones shared on the AIFW official Instagram account.

But now, virtual models are slowly entering the fashion scene. In 2017, the world witnessed the birth of the first digital supermodel: Shudu Gram.

The Supermodel, with beautiful dark skin, majestic height and facial features that scream perfection, is the creation of fashion photographer and founder of the Diigitals agency Cameron-James Wilson.
She's gracious, charismatic and portrays the creations of some renowned fashion brands even though she doesn't exist in real life.

She's the result of computer-generated work; allying fantasy with reality.

In a span of 6 years, the virtual model collaborated with fashion houses like Balmain, Louis Vuitton and Karl Lagerfeld. She's been featured in Harper's BAZAAR and graced her first red carpet at the 2019 BAFTA Film Awards.

The supermodel has amassed a portfolio that looks like one of a real model. WILD!

The online community is split between some seeing it as an opportunity to connect with the new generation interested in technology, and others deeming it as a risky move that could push models out the door. One thing is undeniable, the virtual supermodel continues to secure gigs.

"I’ve never felt more empowered to be able to create what I see in my head. It’s so incredibly fast for me to just communicate to somebody what I’m thinking visually in my mind." Cameron told Country and Town House magazine.

After the success of Shudu, the emergence of CGI and AI-generated models became prevalent in social media and fashion advertisements. The Diigitals agency has expanded its portfolio with 5 additional diverse characters, and it appears that they have no intention of slowing down.

As of today, the agency counts Shudu, Dagny, J-Yung, Brenn, Koffi, Galaxia and three other virtual ambassadors; including Kami, the world’s first virtual avatar with Down Syndrome.

In 2023, another virtual persona surfaced on the internet, this time it was Spain's first AI model.

Looking at the pictures shared by her creator, it was hard to believe that the Aitana Lopez is not real.

The 25-year-old virtual model lives in Barcelona, she has a rosy pout, pink hair and impeccably filled eyebrows. To this day, she's got over 300k followers on Instagram- 300k online users intrigued by her lifestyle and day-to-day life.

In an interaction with Euronews, her creator Ruben Cruz mentioned that she makes up to 10 000€ with an average of 3000€ per month. Since its creation, Aitana has caught the eye of many brands and currently serves as the face of a sports supplements company.

Cruz and his agency crafted an entire character for her: from the way she dresses to her lifestyle and passions. After analysing trends, the agency made her resonate perfectly with the current preferences of oriental culture and contemporary society, hence the pink hair and the interest in fitness and games.

Yet, it's important to acknowledge that while her flawless appearance sparks admiration and fascination, it also raises concern about its potential impact on social perception of body image.

With the AI-generated models becoming more and more refined through computer programs, it became difficult to tell the difference between an AI-generated model and a real human. But above all, one begins to ask specific questions. Are these virtual creations disrupting the industry's traditional workflow and essence? Are we witnessing a scenario where virtual models are taking over real ones? Are real models affected in any way by this revolution? The answers to all these questions are quite elusive and vary from one person to another.

Michael Musando shared his point of view about the use of AI-generated models in fashion, he said that they will never be as impactful as the real ones. “Real models can create interest in consumers, you need real models that can have social influence and create traffic towards websites. This is something AI-generated models can't do. These models are generated so they’re not resembling anyone, so it’s not one of the other it’s a supplementation point of view.” he explained.

On the other hand, Alexsandrah Gondora -a London based model- who met Cameron a few years ago explained during an interview that the digital revolution is so exciting and she doesn't feel threatened by it.

She landed the job of Muse, portraying Shudu for an editorial for Tiffany X Vogue Australia. However, she didn't only become Shudu's muse, Alexandrah owns a virtual version of herself for the purpose of getting more modelling work.

An initiative that Madhulika Sharma—an Indian model living in New York—is open to, with some conditions.