Over the past year, artificial intelligence finally began living up to its name, thanks to a newer, smarter generation of machines that is all the rage for the technology and business sectors. The tech has been headlining one earnings report after another, stoking an AI arms race among major players such as Google, Microsoft and Meta.
Where it goes from here, especially for fashion and retail, can extend far beyond the mere customer service chatbot. How far will depend on the brand, but according to Azita Martin, vice president of AI for retail and consumer packaged goods at Nvidia, many are eager to explore their breadth of options.
Related Articles
“Most retailers come to Nvidia [asking] about the best generative AI use cases that will give them the biggest bang for the buck. We’re certainly seeing use cases that leverage large language models [LLMs],” Martin told WWD, referring to deep learning algorithms that take in vast volumes of data to create surprisingly human-like understanding and interactions.
You May Also Like
The chipmaker, already a giant in gaming, is also one of the most buzzworthy names in AI. The company is fresh from reporting a revenue haul of $13.5 billion in its second fiscal quarter last week, thanks to historic demand for its generative AI processors.
Martin notes that retailers have been using bots as product advisers to recommend items, but she has also been working with several to combine LLMs with customer account data to create more finely tuned AI sales associates. These bots would have a better handle on the shopper’s preferences, once they log into an e-commerce site and identify themselves. Another trend she has been noticing is the use of AI to help write product descriptions. Machine-made content is one of the highlighted features of AI, specifically generative AI, which understands natural language prompts and can generate text, images, audio and more based on those cues.
Nvidia introduced its latest tools at the Siggraph graphics conference in Los Angeles earlier this month. “[We] announced the ability to use images, videos and even photorealistic 3D rendering of product as part of generative AI, and being able to use those in a way that complies with licensing requirements,” explained Nvidia’s Martin. “So, for example, we work with Getty, Shutterstock and Adobe, and our Picasso Foundation Model can be trained with their data, which is primarily images. Then you can basically say, ‘Show me my product with different backgrounds’ — like sitting on a table, with the parents [in the] backdrop and things like that.”
AI tools appear to be maturing at a rapid clip and with business uses in mind. From internal creative to customer-facing interactions, there are few areas of the business where AI couldn’t apply.
We’re seeing more and more briefs from the large brands and retailers that are specifically asking for the use of AI as a tool that they could give or enable consumers to leverage in an immersive experience.”
Tyler Moebius, SmartMedia Technologies
Alice Delahunt, founder and chief executive officer of digital luxury platform Syky, believes that “generative AI will profoundly impact the luxury fashion industry, spanning the end-to-end fashion process.” AI’s efficiencies could reduce human labor in some areas, while expanding the need in others. Executives should consider “upskilling or reskilling their workforce in ways that complement generative AI,” she added. Training staff on an image generator like Midjourney, for instance, can drive creative storytelling by accelerating the concepting for campaign shoots or even creating the assets based on prompts.
AI’s own capabilities run deep, but one of its most dynamic areas is the interplay with other technologies, such as Web 3.0 — which looks like a natural fit to Tyler Moebius, CEO of SmartMedia Technologies, a loyalty platform that has worked with brands including Burberry, Lacoste and many others. “One of the things to watch is how AI comes together with blockchain,” he said.
Luxury brands and designers concerned about proving provenance, ownership and scarcity for products can rely on blockchain, a decentralized technology that’s almost impossible to fake. Right now, AI has its own challenges with ownership rights and intellectual property — the tech is a notable negotiation point in Hollywood’s writers strike, and the use of copyrighted books to train AI models has sparked lawsuits from authors, including Sarah Silverman. There’s still plenty to sort out, but once an ownership, authorship or recognition structure is settled, Moebius views “smart contracts from Web 3.0 and the blockchain being perfect tools” for executing revenue splits, royalties and the like.
Moebius expects to see other types of experimentation as well, including new consumer experiences. “We’re seeing more and more briefs from the large brands and retailers that are specifically asking for the use of AI as a tool that they could give or enable consumers to leverage in an immersive experience,” he said.
As an example, he described a loyalty program his firm launched recently for X Games, which offered custom AI-generated NFTs as fan club membership cards. Upon joining, people provide prompts that are unique to them and the AI uses that to generate a 3D membership card, which is minted as an NFT to unlock perks.
A new virtual destination in Decentraland will cover similar themes: The AI World Fair will be held in the new Innovation City area starting on Oct. 19. The three-day event, produced by Web 3.0 creative agency Verse Digital, will “showcase AI for metaverse products and practitioners.”
Some digital fashion companies are already driving the next AI push: The Fabricant is building its own AI image generator — or, more specifically, a tool that can take images of people and visuals of their virtual fashion, and then generate new photorealistic images with the subject dressed in the digital finery. Its model was trained across a variety of human poses and garments, so the system can create believable pics.
In a test set of 16 generated images shown confidentially to WWD — though not supplied for this story, because the unreleased tech is in development — the quality appeared impressively realistic. Few observers might guess that the suit on the male model is made of code, not fabric. The work is The Fabricant’s answer to the most common questions it receives, which is from consumers asking what they can do with digital fashion and how to wear it.
“The use cases I see for this is social media content, music, videos, fashion shows — basically, any content you can think of,” Kerry Murphy, cofounder and CEO of The Fabricant, told WWD. “But how does this work for us? Well, we get to dress our community, our users and we get to start making fashion runway shows of videos that never existed using our community, our collection holders, essentially.”
Murphy can also envision how his tech could spotlight real-world fashion. “One thing that’s going to happen is ‘see now, wear now,’” he continued. “You see an item on the runway, and you’re like, ‘Oh my God, I love that item. Let me dress myself wearing that right now.’ So that’s one trick that’s going to come from this.”
The Fabricant, which recently created its first digital couture collection, Primal Rave, will take its expertise in digital fashion and its AI ambitions directly to Nvidia next, when it joins the cohort for the tech giant’s Inception program for startups.