TEXTILES.ORG
A lot of uncertainty surrounds artificial intelligence (AI). What are its limitations? How many human jobs will it replace? Can it be trusted with our private information and security? Will it eventually outsmart us and take over the world? The answers to those questions are still unknown, but one thing is certain: Some businesses in the textile industry are using AI to create truly impressive results.
AI and the quest for zero waste
Shelly Xu, CEO and founder of SXD in New York City, N.Y., recognized a huge problem in the apparel industry. “Today’s fashion design process is like buying completely new ingredients every time you cook and leaving what is unused to waste,” Xu says. “The result is that one garbage truck’s worth of fabric is dumped into a landfill or burned every second. Rather than having unconstrained designs that do not account for the limited resources on our planet, I wanted to maximize design creativity within constraints, so we can respect what we have.”
Xu, who has been creating zero-waste designs for 15 years, turned to AI to attack the problem of wasted fabric in the apparel industry.
“I started applying AI to SXD’s work when I realized that even the biggest fashion or apparel companies in the world only have 1–2% of the market,” Xu says. “This means to really move the needle on waste removal in our industry, we need a scalable platform that can work across brands, that anyone in the industry can use to convert their products to zero waste.”
Xu partnered with a world-renowned AI professor from Tokyo and an Adobe researcher from London to build the AI algorithms necessary to create that scalable platform: SXD AI. Not only has it helped realize the goal of dramatically reducing waste, but it also provided other advantages, such as saving time.
“With our AI platform, zero-waste designs can be scaled across fabrics, sizes and similar styles as ‘living patterns’ without fabric waste. This is exciting because zero-waste designs have existed for a long time—for example, the kimono from Japan or sari from India. But it hasn’t been scalable because every time any variable changes, even just a little, a designer would need to manually redesign everything. But with our AI platform, this redesign process can now be done in seconds rather than weeks or months.”
And saving time and reducing waste means greater savings for SXD’s partners. “When we turn our partners’ garments into zero-waste designs, we see up to 46% reduction in fabric consumption. Since fabric tends to be the biggest cost in making garments, this is a significant cost saver in addition to faster scalability,” Xu says.
All this experience with AI has given Xu insight into its strengths as well as areas where humans are indispensable—at least for now. “Great fashion or apparel designs reflect deep understanding of fabrics and clever ways of making garments that feel fresh without compromising fit. These are not things AI can easily replace. Instead, what AI technologies can do is to bring design visions to life in the most efficient way that does not waste materials and help scale that vision for more wearers, democratizing great design while respecting the limited resources on our planet.”
The perfect fit: AI and tailoring
For Bold Metrics, a San Francisco, Calif.-based company, AI forms the backbone of its mission to produce accurate body dimensions and consumer sizing recommendations for clients such as Burton, Columbia, Men’s Wearhouse and Vuori®. The company uses AI to determine more than 50 body data measurements to make a digital twin of each user.
Daina Burnes, co-founder and CEO of Bold Metrics, says the company recognizes using AI as a powerful approach that can tackle the problems of fit, sizing and high return rates. “We leverage AI for predictive modeling and data analytics to generate accurate body dimensions and subsequent consumer sizing recommendations. This includes building AI technology that generates detailed body measurements, surfaces sizing recommendations within the context of fit preferences, and analyzes garment specifications to further personalize sizing solutions by product.”
By harnessing AI, Bold Metrics has realized time and cost savings for the business and its clients. “Our solutions help our clients streamline operations, optimize design workflows and make more informed decisions based on data-driven insights,” Burnes says. “AI has enabled us to scale our operations more efficiently, allowing us to handle larger volumes of data and deliver personalized solutions to our clients.”
Burnes was optimistic about AI, and it has exceeded expectations. “What has been particularly surprising about the capabilities of AI technology is its ability to evolve and adapt to new challenges and scenarios continuously,” Burnes notes. “I also continuously find myself in awe of the incredible accuracy of our AI digital-twin technology in determining more than 50 detailed body measurements of consumers.”
Putting artificial intelligence to the test
While companies such as SXD and Bold Metrics are putting AI to practical use, researchers are continuing to investigate the advantages and limitations of AI. One of these researchers is Seonyoung Youn, who recently received a Ph.D. in fiber and polymer science from Wilson College of Textiles at North Carolina State University in Raleigh, N.C.
“My initial interest in AI began with 3D garment simulation programs, commonly referred to as 3DGS or CAD [computer-aided design] programs,” Youn said. “These tools are very popular with global apparel companies such as Adidas, Nike, H&M and Under Armour. They can significantly save manufacturing time and costs by simulating product designs in a virtual environment before producing physical samples.”
Youn conducted this research funded by the North Carolina Defense Manufacturing Community Support Program and collaboratively worked with Andre West, Ph.D., and Kavita Mathur, Ph.D. She wanted to test whether the accuracy of AI was sufficient to benefit the apparel industry. “Specifically, I’ve explored whether the commercially available AI-based textile digitization technology can accurately predict textiles’ mechanical and physical properties from scanned images and whether it can authentically represent the way textiles behave in a virtual environment,” Youn says.
To do so, she devised an experiment in which she compared conventional digitization and AI digitization methods to evaluate those physical properties. She discovered that conventional digitization required about 20 minutes to analyze a fabric sample, while AI took about five minutes per sample.
Youn also found that the AI-based textile digitization method performed well in terms of fabric’s visual simulation. However, AI did not perform as effectively in representing garment stress on digital twins. “I observed a noticeable discrepancy between the traditional digitization methods and the AI-based methods in this area,” she says. “This gap highlights a critical area where AI-powered textile digitization still needs refinement to fully emulate the mechanical behaviors of textiles, particularly under conditions of stress or strain. Accurate mechanical representation is crucial for predicting performance, enhancing design
capabilities, optimizing materials, improving cost and time efficiency, ensuring wearer’s comfort, and fostering innovation and research in textile engineering.”
Still, Youn thinks AI has a bright future in the textile industry. “While the current technology may not yet be accurate enough to completely replace physical property testing, especially for garment stress analysis, it offers a fast and easy method for the initial design and visual testing phases,” she says.
Will AI replace humans?
So impressive is the ability of AI to “learn” new skills that it’s hard not to wonder what its ultimate impact on the workforce will be. There’s no question it will replace a percentage of human employees, while many other workers will be using or interfacing with an AI system in some capacity in the very near future. Those looking to advance their careers might want to familiarize themselves with AI as more employers like Burnes adopt this new technology.
“Overall, AI skills are an important factor that we consider when assessing the suitability of candidates for roles within our company,” Burnes said.
Youn believes humans will still be needed in the textile industry, although she predicts there will be changes. “While AI does automate certain tasks—particularly those that are repetitive and labor-intensive—it does not necessarily mean that AI will replace all workers in the textile industry. The transition brought about by AI will likely shift the types of jobs available, requiring a workforce that is adaptable and skilled in new technologies.”
The ultimate impact that AI will have on the workforce remains to be seen. However, in the textile industry, innovative businesses are already capitalizing on AI capabilities and making plans to further utilize its potential.
Alan Pierce is a freelance writer based in Minneapolis-St. Paul, Minn.
SIDEBAR: Artificial intelligence comes of age
Artificial intelligence seems inescapable these days and is likely to become more prevalent.
“The methods are here to stay, and they’re going to change a lot of things in particular—a lot of jobs, the way people work and, of course, they’re going to change education,” says Maria Gini, professor of computer science and engineering at the University of Minnesota, where AI is one of her specialties.
“I’ve been in AI since the very early days, which means I followed the dreams, ideas, the disappointments,” Gini notes.
While AI is clearly the next big disruptive technology, she doesn’t think the human touch is hurtling toward obsolescence. “I think in the AI community, there’s more and more interest now in thinking about the human/AI collaboration. So it’s not that AI is just an independent entity that makes all the decisions, but it’s a tool for humans.”
Gini, who is also an expert in robotics, sees strong comparisons between the use of robotics and the use of AI. Robotics, she notes, has taken over tasks such as painting and welding to free up people for more creative roles. And just as robotics has taken away jobs, it has also created new work for people. She thinks the lessons of robotics apply to AI.
“If you assume that AI is going to replace people so you don’t need workers because AI will do everything, you’re going to be wrong,” she says.