Fashion meets AI: The next step in material innovation The - Première Vision Paris - Denim Première Vision - Première Vision New York
Materials, textiles and leather are at the heart of fashion. They define its very essence: the volume of a garment, its feel, aspect and overall character is intrinsically linked to the material, whose structure, texture, colour, touch and behaviour embody its desirability.
And textiles are among the most complex of materials. The myriad characteristics of fibres, along with their preparation, yarn assembly, structural combinations, treatments, finishes, and printing techniques, create virtually limitless possibilities. Each combination gives a fabric its unique identity, and the ability to craft the perfect blend of characteristics is rooted in the deep expertise and time-honored craftsmanship of textile designers, weavers, knitters, and manufacturers.
Yet, this complexity does contribute to significant environmental challenges. Blends of fibres and certain structures hinder recyclability, while textile production (including dyeing and finishing) accounts for over 50% of the apparel sector’s greenhouse gas (GHG) emissions. While the different raw material extraction or production stages are inequal in terms of resource depletion (land, water and energy use ) and GHG emissions, weaving, dyeing, and finishing processes are particularly impactful, often exacerbated by insufficient treatment of toxic chemicals and wastewater discharge.
In response to increasing legislative pressure—such as the Corporate Sustainability Due Diligence Directive (CSDDD) and the Responsible Textile Recovery Act—brands are now obligated to delve deeper into their supply chains. Designers must reconsider their materials, gain interest and understand compositions, as well as production methods while navigating issues around supply, pricing, and quality.
AI is emerging as a crucial tool to address these challenges, offering advancements in resource efficiency, waste reduction, clean chemistry, material innovation, and recycling techniques. Let’s explore how AI is transforming material development in these key areas.
Read also: Eco-question: Corporate Sustainable Directives – what are the differences between CSRD and CSDDD
Resource efficiency: Optimising material use through intelligent systems
Sustainability, as defined by the Brundtland Report (1987), is « development that meets the needs of the present without compromising future generations. » AI technologies are aligning with this principle, offering ways to reduce resource consumption and waste while maximising material durability and quality, which is key to achieving SDG 12 (Responsible Consumption and Production), one of the seventeen UN Sustainable Development Goals (SDGs).
First, in terms of energy, water and chemistry use, but also in terms of natural resources, particularly through precision agriculture. Utilising data from sensors, satellite imagery, and weather forecasts to optimise water usage, fertiliser application, and pest control.
Additionally, AI-driven yield prediction models analyse historical and real-time data to forecast crop output, allowing for better resource planning and preventing overproduction. By leveraging APIs, these systems integrate diverse data sources, enabling real-time decision-making that reduces waste and lowers environmental impacts in raw material production. These technologies are already being employed by EUCOTTON regenerative cotton suppliers in Spain to reduce water use and enhance production efficiency.
Leveraging digital technologies to track, analyse, and optimise the use of sustainable bio-based feedstocks such as their local or seasonal availability, enables the emergence of competitive markets and dynamic pricing for such resources.
Textiles of the Future, ETP Partnership under Horizon Europe
Waste reduction: leveraging Data-driven insights to minimise unused materials
AI can optimise raw material use right from the extraction stage by precision blending which consists in analysing the characteristics of different fibres to suggest optimal blends and improve the quality of the yarn while reducing waste from rejected materials. Another axis is raw material allocation: AI models predict the required amount of raw fibres needed to meet production goals, helping manufacturers avoid over-ordering and wasting materials.
Companies like Smartex and Oshima use AI to identify defects in fabrics, reducing waste by stopping production when issues are found. AQC, partially owned by Lectra, provides automated, cloud-based quality control that analyses fabric rolls at high speed, ensuring minimal material loss.
This real-time optimisation extends beyond yarn production to spinning, weaving, and knitting. AI adjusts these processes based on real-time conditions, assuring uniform quality and reducing material waste. It also can optimise processes, adjusting speeds and tension based on real-time conditions, ensuring uniform quality, reducing energy and material waste.
We will be considering the waste management in garment production itself more in depth in the next article dedicated to manufacturing.
Increased circularity: AI-powered recycling techniques
AI is also playing a role in post-production waste reduction, addressing the big challenge of overproduction and helping the fashion industry move towards a circular economy, where textiles can be repurposed rather than ending up in landfills.
From the robotisation for sorting to material identification and blend separation, AI enables recycling and a potential reuse of the fibres, with increased speed and efficiency. The NIR (near-infrared) sensors and cameras are one of the key tools, as used by Newretex e.g., with robots sorting textile waste based on material composition and colour identification.
The composition then can be analysed via the NIR spectroscopy, a rapid and non-destructive optical technology analysing the molecular structure of material. However, challenges remain, especially with multi-layered fabrics or textiles with coatings and prints, as current NIR sensors only detect the outermost layer, or loosely knit garments which may also be categorised incorrectly, as the sensor struggles with recognition.
Tackling the issue of mixed fibre fabrics, Epoch Biodesign enables infinite recycling of polyamide blended fabrics into new chemical building blocks that can be used to manufacture new, recycled Nylon materials. They train their generative Al on the language of biology to design novel enzymes capable of transforming complex pre and post-consumer polyamide (nylon) waste with no other viable recycling option, from elastane-blended sportswear to high-performance multi-layer laminated waterproofs.
Read also: Smart Key: Recycling—can the fashion sector become self-sufficient in resources?
R&D of new materials: Accelerating innovation of fossil-free alternatives
In regard to the unpredictability of climate, geopolitical tensions and global uncertainty, the diversification of fibres is becoming critical not only to mitigate risks, but essentially in regard to the urgency to preserve natural resources, giving nature the time to regenerate by stopping deforestation, ( EUDR deforestation regulation), intensive monocultures, leading a loss of biodiversity and every year, and earlier « Earth’s overshoot » day, where human’s demand on nature exceeds Earth’s biocapacity.
Alongside certified organic and regenerative agricultural practices, responsible sourcing of plant-based materials and recycling technologies, AI is accelerating the development of green chemistry, next-generation biomaterials and biosynthetics, offering alternatives to conventional, fossil-based fibres and products.
Historically, innovation in textile fibres took place over long periods of experimentation. Today, artificial intelligence makes it possible to compress this time by rapidly analysing thousands of combinations of materials and growing conditions, fusing bio-manufacturing, advanced protein engineering, and molecular biology.
Latest innovators include Solena with their AI-based protein design for novel fibres, or Nanoloom, using algorithms to intelligently select which fibres to combine, understand the perfect ratios for each blend, and predict the exact characteristics of the resulting fabric or knit. Nanoloom’s AI systems manipulate graphene nanomaterials and their unique properties, to enhance fibre strength, conductivity, and recovery.
Finnish company Spinnova, known for their textile fibre made from wood pulp or other waste, that is mechanically refined and transformed into spinning-ready fibre suspension without harmful chemistry nor dissolving process, employs AI to optimise various aspects of its operations, including material development and production efficiency.
AI-based biotechnology for colour chemistry virtuous solutions are also at rise, we’ll go into the details in a following article.
Read also: Smart Key: How to color garments without leaving a stain on the environment?
3D simulation: AI-enhanced material digitisation
Finally, one of the biggest tools to avoid overproduction, unleashing also design creativity, is 3D visualisation of fabrics. Companies like Vizoo, (3D scanner and software for photorealistic 3D visualization), VSK Technology, (whose NAO Virtual Knitting Machine can realise 100 different fabric constructions per second and display these on screen in real time) and Hohenstein Fitting Lab who showcase digitisation of materials that can directly be integrated into 3D simulation tools like CLO3D, Browzwear and Vidya from Style3D that are at the forefront of digital simulation.
The expertise on material simulation is also closely linked to colour management, and although both remain a high challenge to be translated in physical reality, platforms like ColorDigital GmbH’s DMix are working to enhance precision in digital texture mapping, to capture the subtlest details and nuances of a wide range of materials.
Yet the magic of material – matter-real, real matter – lies in the surprise of the tactile qualities, weight, softness or roughness, that are not what they may look like. As the French saying goes, “L’habit ne fait pas le moine” (the clothes do not make the man), visuals still are not what a fabric’s essence is.
To conclude, although these AI based solutions are hope bearing, as we see the current going, the global textile production is not going to stop its growth. It has nearly doubled in the last two decades, and what is even more alarming, is that synthetics, especially polyester, constitute now 67%, polyester 57% of the global production. That’s why “under the business-as-usual ( which we are back to) growth projections, GHG emissions of the apparel sector will grow to 1.588 Gt by 2030, well off pace to deliver the 45% absolute reduction needed across all sectors to limit warming to the Paris Agreement’s goal.”*
Ultimately, AI must be used strategically to meet genuine needs and encourage thoughtful decision-making. While AI can optimise processes and generate data-driven insights, its fastness should be at the service of good purposes, enabling humans to slow down, clear their vision, and leaving space for real innovation – research for novel, unexpected, unpredictable surprises that incarnate the beauty of human creativity.
OTHER ARTICLES FROM THE FASHION MEETS AI SERIES
*https://www.wri.org/technical-perspectives/roadmap-net-zero-emissions-apparel-sector