The AI image generator Dall-E Mini is addictive. Type a description of anything you can think of – a bottle ranch dressing testifies in courtBoris Johnson entangled in a net under water† Eminem as M&M – and it will bring up images that match each other.
And when you start giving it the names of fashion designers, this ability to distill a text prompt into the picture is fully displayed. “Rick Owens” returns human forms draped in black garments that Owens could have easily imagined on his own. “Dries Van Noten” delivers Dall-E Mini’s interpretation of models on a catwalk, dressed in a vibrant, Van Noten-esque explosion of prints and colors.
The AI is effective not only in identifying a brand’s key shades, but also silhouettes and other signatures. For example, with ‘Thom Browne’ you get a grid of shrunken gray coats, embellished with stripes.
Dall-E Mini creates images based on his understanding of the relationship between words and pixels by analyzing online images and how they are labeled. It’s not perfect. The images are slightly distorted and smudged, adding to the weirdness of the deliberately strange scenes people have have compiled† It is also limited by the images it has analyzed, so if it doesn’t have a large dataset to draw on the resulting images, it can be reductive. (In my tests, “Saint Laurent” usually means black leather jackets.)
Yet it has an intriguing ability to capture the essence of a designer or brand.
AI remains a far from becoming the conscious machine sci-fi writers have thought. It is adept at finding patterns in large amounts of data and in fashion it is mostly used as a tool for business tasks such as determine pricesto predict customer’s lifetime valueforecasting demand and quick reorder from hot sellers† On the creative side, however, it has had less impact.
Dall-E Mini and other experiments suggest that one day this could change.
Dall-E Mini’s images are mostly variations on what’s already out there, rather than fresh new ideas. But the ability to mimic a designer caught the eye of online retailer Ssense, which: left his Instagram followers guessing the brand inspiration in a series of images shot with Dall-E Mini, as well as fashion writer Derek Guy, who visits @dieworkwear on Twitter and lets me play with the AI for the first time.
There’s also a more powerful model, Dall-E 2 – the successor to the original AI launched in 2021 – that the developers have made accessible only to certain users. The images look much more realistic, making the risk of problematic uses such as deepfakes, AI-generated images and videos so compelling that they are difficult to identify as fraudulent. Whether or not it can generate credible imitations of a designer’s work isn’t entirely clear, but given the way AI is progressing, it could be a matter of time before the scenario becomes a reality.
This situation can have a number of consequences if it ever happens. It’s not hard to imagine a fast-fashion retailer like Shein using AI to generate its designs. Shein already relies on her suppliers to provide clothing designs from which her internal buyers then choose, according to reports in Chinese-language media† It could potentially use an AI model to constantly come up with new variations of clothing for its buyers to choose from. Ultimately, even selecting the designs could theoretically be done with AI based on what’s already being sold, costs, weather forecasts, trend data, and other inputs.
There are ongoing discussions as to whether this would be design at all or represent the further descent from fashion into commodities. Some factory design farms are already producing ideas with little originality or clearly copying the work of others. Is it worse when a machine does it? The work of the AI would still be based on human creativity as it would produce its new concepts based on examples from the past. Although it can also shamelessly copy existing designs, but so can people.
However, AI could open new creative avenues for designers themselves. An artist has granted access to the more powerful Dall-E 2 model told Wired he used it to test more new ideas and speed up creation method. One fashion designer I can imagine using this type of AI to enhance, not replace, her output is Iris Van Herpen, who already relies on technology to design her alien garments. Someone along the lines of Virgil Abloh, sometimes referred to as a curator as much as a designer, could find other creative ways to deploy AI.
Levi’s already sees AI as a promising way to support and complement its designers, Katia Walsh, who leads the company’s AI efforts, told me in a recent conversation. Last year, Levi’s design coordinator Ron Pritipaul devised an algorithm that produced a trucker jacket with a print derived from Vincent van Gogh’s “The Starry Night.” While it’s not the most revolutionary product, it shows how AI can “bring ideas that a designer might not have thought of in the traditional process.” Levi’s said in a blog post at the time. When we spoke, Walsh said she had a dress made from a Starry Night print in her office.
While Levi’s isn’t rolling out AI-designed items yet, the technology is proving useful to the company in more mundane ways. Pritipaul devised AI for color matching threads, where a computer performed a normally tedious and time-consuming process in seconds. Walsh said this kind of application of AI frees up designers to do other work.
Some critics may disagree with the use of AI in design, worrying that it will reduce creativity to mere computing or that it will doom us to rekindle old ideas. But AI is just a tool like any other technology. It is up to man to decide how to use it, and the potential use continues to grow.