Artificial Intelligence for the Fashion and Retail Industry: Insights From Network Analysis of the Current Literature

Abstract

language (Version 3) was used for algorithm development. CONCOR analysis and network visualization were conducted using UCINET for Windows (Borgatti et al., 2002). ResultsThe number of publications per year grew exponentially with the highest number of publications (n=114) in 2020 and the USA was the most productive country (n=63). Keyword frequency analysis showed that research has been most active in the area of consumer ( 408) usage (375) and design/development (86) of products (143) and services(259), especially in the context of chatbot (67) and robots (75). This finding is corroborated by the results of the network analysis. An examination of closeness centrality, which identifies central words in the network (i.e., core keywords that relate to all other words in the network), revealed that keywords such as 'chatbot,' 'robot,' 'fashion,' 'device,' and 'algorithm' were of high importance. Together, it can be concluded that a large proportion of research has been on the development of algorithms of chatbots and robots to facilitate consumer purchase of fashion products. The importance of keywords in the overall network of themes was examined through betweenness centrality of a node, the number of times the node is included in the shortest paths of any pair of nodes in the keyword network (Radhakrishnan et al., 2017). The most diagnostic keyword among the top 30 keywords was "social". Further examinations of the literature revealed that social was an important word in service/product development and in the innovation acceptance process of consumers. Phrases such as 'social-robot,' 'social-service,' 'social-oriented,' and 'social-role,' reveal that the development of AI-enabled services/products with social functions is actively progressing. In addition, through words such as 'social-consideration,' 'social-perception,' 'social-influence,' and 'social-contextual,' it was found that social is an important factor when AI-related products or services are evaluated by consumers. Then, CONCOR analysis was conducted to reveal concurrently appearing words because this technique can reveal hidden subgroups and the semantic structure of text (Breiger et al., 1975). Four themes emerged from the analysis: 'System development & application,' 'Perception of robots,' 'Role of AI,' and 'Healthcare.' The greatest number of articles on AI dealt with the system development and application. Popular keywords ('machine-learning,' 'management,' 'algorithm,' 'system,' 'application,' 'design') suggest that attention has been given to the development of AI-based systems and applications. Another notable topic was consumer perception of robots and chatbots ('social', 'robot', 'trust', 'chatbot', 'perceived.'), suggesting the research interests in the level of consumer trust in AI. Meanwhile, words such as 'consumer,' 'market,' 'service,' 'marketing,' 'role, ' 'retail,' 'fashion,' and 'IoT,' indicate that research on the diverse role of AI creates another stream of research. Lastly, the heightened anticipation and excitement for AI-enab...