This page was printed from

AI’s potential in saving manufacturers time and money

Policymakers, manufacturers, computer experts and early adopters need to collaborate.

Markets | June 1, 2024 | By: Seshadri Ramkumar, Ph.D.

Smartex, an automation company in Porto, Portugal, produces AI-enabled products for textile factories, such as the Smartex CORE pictured here. Smartex CORE is designed for circular knitting machines and uses cameras and sensors to scan the fabric for irregularities and faults. It also uses light analysis to catch defects not visible to the human eye. Image: Smartex

Artificial Intelligence (AI) has gained global prominence since serious discussions on this subject surfaced at the 2017 World Economic Forum in Davos, Switzerland. Following the 1990s’ information technology (IT) revolution, we now have arrived at the point of harnessing the power of machine intelligence coupled with human intellect. According to Kiran Minnasandram, chief technology officer of Cloud Business at Wipro Technologies, Bangalore, India, “The synergy between AI and human intellect pushes the boundaries of what is achievable in manufacturing, leading to products that embody both the precision of automation and the creativity of the human mind.” 

In their book Tools and Weapons: The Promise and the Peril of the Digital Age, Brad Smith, vice chair and president at Microsoft, and Carol Ann Browne, general manager and chief of staff and executive communications at Microsoft, highlight the importance of reliability, transparency, safety and accountability in enabling the acceptability and wider applicability of AI. Concerns have surfaced regarding privacy and intellectual property issues as well as ethics questions, data privacy and social impacts, hence the need for governmental regulation. But AI has also been touted as an enabler of growth. 

According to the U.S. Chamber of Commerce’s AI Commission Report, AI will enable about $13 trillion in economic growth by the end of the decade. While defense technologies were considered a country’s main strength during the Cold War era, soon knowledge-related fields such as AI will take on this role—in addition to serving as an indicator of economic strength. 

Because the speed and capacity of computer processing essentially is the enabler for the development of AI, several factors must be considered when evaluating AI’s output. The quantity and quality of data, standardization of data set collection procedures, repeatability of resources, reliability of the data and its accuracy all affect AI’s effectiveness.

The WiseEye textile inspection technology has been under development at the Laboratory for Artificial Intelligence in Design (AiD Lab). It can detect faults in common types of textiles, with simple patterns and stripes not an issue. It runs at 60 meters (197 feet) per minute and can be adapted to leather. Image: Laboratory for Artificial Intelligence in Design

AI’s adoption in textile sectors

Information technology and machine learning algorithms must be refined to suit the requirements of the field, such as relating cotton yield predictions to fashion markets influenced by consumers’ choices in apparel and fast fashion. In fact, among the subsectors in the textile industry, design and fast fashion have been leading in the exploration and use of AI.

The apparel sector is seasonal and based on consumer preferences. Understanding it intertwines social sciences, behavioral psychology and manufacturing. AI-enabled machine training of patterns over different seasons allows brands to develop a variety of designs and styles that may have broader appeal. With the help of multiyear data on fashion trends and consumer preferences, AI tools can help predict the future. The key is to work with the optimum data sets in quantity and quality for reliable predictability.

Fault detection in fabrics using machine vision and image analysis is needed. The reliability of fault detection depends on the availability of large data sets to train the algorithm for good repeatability and reliability.

“Through advanced analytics, machine learning algorithms, and real-time monitoring, AI technologies can identify and rectify defects in the manufacturing process much earlier than traditional methods, significantly reducing waste and ensuring that products meet high-quality standards,” says Minnasandram. 

Supply chain barriers affect costs as well. Mohamed Suhail, an AI researcher and assistant professor in computer applications at National College in Tiruchirappalli, India, explains, “By anticipating demand, maximizing inventory levels, and streamlining logistics and distribution procedures, artificial intelligence can improve supply chain management. As a result, the supply chain will be more efficient and enable cost savings.”

In the textile sector, raw materials represent about 60%–70% of the total cost, so having a good handle on inventory is a cost saver. As fiber availability—particularly in the case of natural fibers—and its price volatility play important roles in influencing the market, proper prediction is necessary. AI-based decision-making and market predictions in volatile sectors, such as cotton, are slowly penetrating the field. The IT sector and the manufacturing sector must interact to standardize this kind of data collection and analysis. 

The advanced textiles sector

The advanced textiles sector has been slower in realizing the potential of AI tools for its development and growth. Because the development of new fabrics and functional chemistries is frequently related to new fibers, AI could be especially useful in this area. The industry has been insisting on the need to develop a portfolio of high-performance fibers that would be used in defense and medical applications. AI can be coupled with genetic engineering to synthesize new fibers with new applications and functionalities.

Worth noting are efforts to incorporate the concept of tandem repeat, used in DNA tracing, with AI and synthetic biology to develop self-healing fibers. For example, by understanding the genetic information of the protein in a squid’s ring teeth inside the animal’s suction cups (as reported in several science publications), biomanufacturing of squid protein at mass levels is possible using bacteria. The required high-performance properties, such as superior elasticity, can be modeled using AI technologies, and they can be tailored to develop sustainable protein-based fibers.

AI can also be used to develop next-generation defense, medical and protective clothing. In these fields, obtaining a balance between protection and comfort is an issue. By examining data from different users, performance and requirements, the optimum functional, physical and chemical attributes can be optimized. 

The future of artificial intelligence

As the advanced textiles sector and the textiles industry in general are just beginning to realize the potential of AI, reports on the benefits of AI in manufacturing are surfacing. Minnasandram says, “Global surveys show that after integrating AI-driven systems, some manufacturers have reported up to a 20% reduction in production costs and a 50% decrease in downtime due to predictive maintenance. AI algorithms are exceptionally good at scheduling maintenance only when needed, preventing unexpected breakdowns and unnecessary checkups.”

Commenting on the future of AI in manufacturing as a researcher in the field, Suhail advises, “It’s crucial to remember that adoptability of AI in manufacturing may differ depending on the sector, size of the organization, customer preferences and requirements. Businesses that successfully adopt and integrate AI technologies are expected to acquire a competitive edge in terms of efficiency, cost-effectiveness and innovation.”

The information technology discipline is growing fast. Because AI depends on the availability of quality information, including reliable and large data sets, collaboration among experts in machine learning and textile manufacturing sectors will help the industry to understand privacy requirements, the amount of data required and standardization of data needs. The textile sector needs to have cross-disciplinary approaches, and forums must be developed where a seamless flow of new technological advancements can be exchanged with the manufacturing and textile sectors. Industry organizations, such as the Advanced Textiles Association, need to establish working groups to look at the merits, demerits and the effect of regulations on AI technologies. More important, the immediate need is to engage in broader outreach among policymakers, manufacturers and those already incorporating the technology in their workplace. 

Seshadri Ramkumar, Ph.D., is a professor in the Department of Environmental Toxicology and the Institute of Environmental and Human Health at Texas Tech University.  

Share this Story