Threads of tomorrow: Reddit’s conversation on generative AI across the fashion domain

Generative AI is spreading across the fashion industry, yet public discourse about these tools remains understudied. We analyze 15,854 Reddit posts (January 2022–June 2025) using BERTopic modelling and transformer-based sentiment analysis. Through human topic screening, we identify nine fashion-related themes containing 4836 documents of theme-level analysis. Apparel design generates the most discussion and one of the highest shares of negative sentiment, revealing a ‘democratization paradox’: AI expands access to design but still requires professional refinement. Sentiment also differs by application type, with instrumental uses, such as virtual try-on and personal styling, attracting significantly less negativity than representational uses, including avatars and AI personas. Ethics discussions are dominated by AI undressing tools, a concentrated harm pattern that emerged nearly two years before the TAKE IT DOWN Act. These findings extend technology acceptance frameworks and demonstrate how social media discourse can provide early warnings of emerging risks while guiding fashion practitioners and policymakers.