AI for the fashion industry - What’s next?
2023 was the year of AI innovation. We've seen a significant surge in AI startups reshaping various industries, so what about fashion, one of the world’s largest industries? Let's explore how AI is used in fashion.
Written by Aya Bochman | January 10, 2024
After generative AI’s breakout year in 2023, use cases are emerging across creative industries, including fashion. - BOF - The State of Fashion 2024
How AI is used in fashion so far
AI has completely changed the way we approach innovation in every sector today. If we consider the fashion industry, numerous advancements have already been made possible in terms of creativity, sustainability, and marketing. Let’s break them down:
Creativity
Designers are using apps like Midjourney, which are capable of generating beautiful and realistic high-quality images from their imagination or based on existing images. This technology enables anyone, from individual creators to small and large brands, to create unique collections, concept art and campaigns. Take a look at MaisonMeta for example. This NYC-based Generative AI creative studio utilizes AI across multiple sectors, including fashion, and has already created various concepts and campaigns for major fashion brands such as Moncler and Revolve. Here’s a glimpse of what they achieved using AI for a Revolve campaign:
Sustainability
The advancement of technology plays a crucial role in addressing sustainability, a major concern in the fashion industry. According to Earth.org, 92 million tonnes of textiles waste is produced every year, and these numbers are only expected to rise if innovative solutions are not found.
One approach being adopted by fashion companies involves the use of AI for trend prediction. By utilizing AI models to more accurately forecast trends, brands can produce what is truly needed, thereby reducing unnecessary waste. Additionally, AI can enhance efficient production and optimize supply chains, further diminishing material waste and energy consumption. A notable example is Heuritech, which uses AI to detect thousands of fashion attributes like prints, colors, fabrics and shapes, from analyzing vast amounts of social images.
Trend prediction platform. Source: Heuritech
Marketing
Personalized campaigns have redefined fashion advertising. Much of the content now seen is AI-generated marketing content, tailored to individual consumer behavior and preferences. This results in higher engagement rates and more sales conversions. Additionally, it’s worth noting the emergence of AI stylists that excel at identifying desired items based on images or suggesting unique styles, using AI to create the perfect combination. One example is Style DNA, a mobile app powered by AI technology that helps users discover their personal style, provides shopping suggestions, body fitting, and more.
Personalized recommendations app example. Source: Style DNA homepage
But there is another very important player in marketing: Virtual Try-on. Some brands have already implemented try-ons in their websites with the help of augmented reality (AR), which allows customers to virtually try on items, resulting in an enhanced online shopping experience and reduced returns. It will be intriguing to observe how this technology evolves in the future and whether it will become the standard method for online shopping.
Try-on - The Future of Fashion
We've discussed some incredible AI-driven solutions in the fashion industry, primarily focused on marketing and production. However, one area that is often overlooked, or perhaps not as straightforward to implement, is AI-generated fashion photoshoots.
Consider a major fashion brand like ZARA, which produces at least a dozen collections annually. Each collection requires photoshoots with various models and locations, resulting in a lot of effort and expenses. Before the AI revolution, it was difficult to imagine a solution for such a challenge, But now, with the advent of generative AI, we can come up with solutions like digitally generating multiple clothing styles on a diverse range of models, without the need for actual physical wear or studio visits.
One example of a pioneer in this technology is Google. In June 2023, Google introduced their virtual try-on for apparel, currently available in the U.S. for select brands and for women’s tops only. This innovation is a game-changer for both brands and shoppers, because it allows customers to simulate how a piece of clothing will look on different types of models, as it allows customers to simulate how clothing will look on different models, particularly varying in body types and skin tones, aiding significantly in their purchasing decisions.
Google simulating the same garment on various models. Source: Google virtual try-on
The adoption of try-on technology by fashion brands can profoundly impact the industry in various ways. It's not just about financial savings by cutting costs, but also about reducing environmental waste. This innovation has inspired and motivated our FASHN team to develop our own try-on solution.
“OutfitAnyone” virtual try-on demo from Alibaba’s HumanAIGC. Source: HuggingFace demo
Our Try-on Solution
So, how does our try-on technology work? In brief, the process begins with two inputs: the first is an image of a garment, either worn by a model or displayed standalone, and the second is an image of a chosen model, irrespective of gender, body type, or ethnicity. After submitting both inputs, our platform generates an image of the selected model wearing the garment. This process can be repeated as needed to achieve the desired representation for the garment.
Virtual try-on has the potential to revolutionize how clothes are presented on e-commerce platforms. It can be particularly beneficial for small businesses, which often lack resources and cannot afford to conduct photoshoots with models for each garment. With our try-on technology made accessible to all, fashion brands need only clear images of their garments and images of the models they’ll want representing their brand.
Conclusion
The emergence of AI solutions in fashion, such as try-on technology, marks a transformative moment for the industry. By enabling brands to digitally simulate clothing on diverse models, we are democratizing representation through inclusivity and moving towards more sustainable and cost-efficient marketing practices.
We believe that 2024 will be another fruitful year for AI innovations. However, there is still considerable work to be done before solutions like try-on technology are fully ready for commercial integration. So stay tuned and follow our journey in FASHN!