Repository logo
 

AI-Generated Models and Their Impact on Consumer Trust and Perceived Risk in Online Fashion Retail

Files

Dissertation(3.22 MB)

Date

2026

Authors

Melamed, Milana

Supervisor

Kapitan, Sommer

Item type

Dissertation

Degree name

Bachelor of Business (Honours)

Journal Title

Journal ISSN

Volume Title

Publisher

Auckland University of Technology

Abstract

The purpose of this research is to understand how the use of fashion models generated by artificial intelligence (AI) in product imagery affects consumer trust and perceived risk in the online retail setting. As AI technology continues to advance, its implementation and the opportunities it offers have noticeably expanded across marketing practices. AI-generated models are one of the latest phenomena observed in advertising and retailing spheres. Previous literature has widely explored the effects of different types of sources in marketing content, with recent studies focusing on the use of virtual endorsers and AI imagery in advertising. However, as AI-generated models start to become a more common occurrence in online retail spaces, its impact on consumers in this setting remains under researched. Understanding the effects of this phenomenon is particularly important in the online retail context due to its intangible nature, which can lead to higher perceptions of risk and lower level of trust (Teo & Liu, 2007; Handoyo, 2024). As consumers often rely on visual information to reduce perceived risks, particularly when shopping for clothing online (Yu et al., 2012), the use of AI generated models can be expected to negatively impact consumer trust and perception of risk. Recent studies have demonstrated that disclosed AI imagery in ads is associated with a decrease in credibility and trust (Grigsby et al., 2025), while other studies have shown that virtual influencers have the same effect (Hofeditz et al., 2022; Nissen et al., 2023). Such outcomes can be related to a reduction in perceived anthropomorphism, whereby lower perceptions of anthropomorphism in AI imagery are associated with a decrease in consumer trust (Muniz et al., 2024; Luo et al., 2019). Thus, using a quantitative research approach, the current study aims to address these observed gaps in the literature. This work’s research question seeks to understand how the use of AI-generated models impacts consumer trust and perceived risk, specifically in online fashion retail setting. This research question will be answered using five hypotheses: H1: AI-generated models in online product images negatively impact consumer trust and perceived risk. H2: The impact of AI-generated models on consumer trust and perceived risk is greater when the use of AI is explicitly disclosed compared to when it is not disclosed. H3: The conditions (human model vs. AI model vs. AI model with disclosure) predict varying levels of perceived anthropomorphism H4: Anthropomorphism is associated with consumer trust and perceived risk. H5: Anthropomorphism mediates the effects of model type on consumer trust and perceived risk. Using the online survey platform Qualtrics, the study was distributed through panel recruitment agency ResearchConnect and gathered 202 participants. The research rejects both H1 and H2, however, it fully supports H3, H4, and H5. The study reveals that there are no differences in trust and perceived risk between consumers exposed to AI-generated models (with and without disclosure) and human models in product images. However, the study shows that there are differences in perceived anthropomorphism between the different types of models. The study reveals that perceived anthropomorphism mediates the relationship between model type and responses in trust and perceived risk. This means that when some consumers are exposed to AI models in e-commerce product images, a decrease in perceived anthropomorphism negatively impacts their perceptions of trust and risk.

Description

Keywords

Source

DOI

Publisher's version

Rights statement

Permanent link

http://hdl.handle.net/10292/21054

Collections

Bachelor with Honours Dissertations - open access