Artificial Intelligence

Ten years is a long horizon in any industry, but in AI it may be a category error. Fashion for Good's mapping of a 2036 AI-native supply chain is technically grounded and operationally serious. The structural questions it leaves unanswered, around rebound effects, labour transition, legal accountability, and regulatory reach, are the ones that will determine whether the vision it describes produces genuine progress or faster versions of the same systemic failures.

An AI-driven wardrobe feature has been added to Google Photos, automatically cataloguing clothing items from users' photo libraries into a browsable digital closet. Users can filter garments by category, assemble outfits on digital moodboards, and preview looks via virtual try-on. Developed with Motorola, the feature rolls out on Android from summer 2026, with iOS to follow.

Flaws in large language model-based user simulators have been exposed by research from Google Research introducing ConvApparel, a dataset of more than 4,000 human-AI conversations in the apparel shopping domain. The work develops a three-pillar evaluation framework to measure and address the realism gap in AI testing, with data-driven simulators outperforming prompt-based approaches across key behavioural metrics.

Fashion retail is moving into a sharper phase of algorithmic pricing as AI shopping tools connect consumer preferences, inventory pressures and automated purchasing. Dynamic pricing, virtual try-ons and agentic commerce could make shopping easier, but may also deepen discount chasing, impulse buying and overconsumption in already high-volume markets shaped by markdowns.

IIT Roorkee has developed a new artificial intelligence framework that enables the translation of detailed textual descriptions into professional-quality fashion sketches. Built around the FLORA dataset and the NeRA adapter, the research addresses limitations in existing AI tools and supports more precise, expressive and sustainable design workflows across fashion, retail and allied creative industries.

Researchers at Pusan National University have developed a systematic prompting method to use ChatGPT for predicting fashion trends. Their Top-Down Prompting technique, inspired by the Lotus Blossom approach, helps generate more accurate and consistent insights, potentially democratising trend forecasting for smaller players and fashion education.

Researchers in China have used artificial intelligence to design better carbon-based materials that can clean toxic dyes from industrial wastewater. By analysing hundreds of lab results, the system identified the most effective combinations for dye removal, helping make low-cost, eco-friendly solutions for the textile and chemical industries more efficient.

Artificial intelligence is reshaping agriculture by offering smarter ways to monitor pests and optimise pesticide use. A new initiative demonstrates how rural communities can harness cutting-edge technologies to strengthen farming practices, protect health and build innovation capacity. The project also highlights how collaborative partnerships can bring long-term economic and community benefits outside urban centres.

Korean researchers have demonstrated that generative AI can successfully predict and visualise seasonal fashion trends, achieving accurate implementation in over two-thirds of tested cases. The Pusan National University study used ChatGPT and DALL-E to forecast Fall/Winter 2024 menswear collections, generating images that closely matched real runway designs whilst highlighting the continued importance of fashion expertise.

Virtual influencers — digitally created personas used to provide entertainment, generate content and endorse brands — are becoming increasingly influential, especially among Gen Z and digital-first audiences. The rise of virtual influencers is notably prevalent in Asia.