The Effect of AI Disclosure Formats on Consumer Responses in Fashion E-Commerce
Author
Naveed, Anas
Term
4. Semester
Publication year
2026
Abstract
This thesis examines how different ways of disclosing the use of AI-generated visuals influence consumer responses in fashion e-commerce, in the context of the EU AI Act’s requirement to label AI content. While prior research has largely compared simple disclosure versus non-disclosure, this study focuses on how three specific disclosure formats—static, progressive, and justified—shape trust, perceived brand effort, and purchase intention. Guided by the Persuasion Knowledge Model and Cognitive Load Theory, the research adopts a qualitative exploratory design based on 18 semi-structured interviews with active online fashion shoppers in Europe. Participants were each exposed to one of three video ads using different AI disclosure formats, and visual elicitation was used to capture their cognitive processing and emotional reactions. The findings show that how AI is disclosed matters more than whether it is disclosed at all. Static disclosure creates immediate but moderate skepticism, strongly influenced by prior familiarity with the brand. Progressive disclosure generates the most negative responses, as delayed revelation of AI use is widely perceived as an intentional attempt to hide information. By contrast, justified disclosure presents AI as a human-guided tool, which reduces algorithm aversion, enhances perceived brand effort, and helps maintain purchase intention for most participants. Theoretically, the study extends the Persuasion Knowledge Model by demonstrating that consumers’ awareness of persuasion is not triggered solely by the presence of a disclosure cue, but also by its timing, framing, and contextual justification. Practically, the thesis offers fashion e-commerce firms actionable guidance on how to comply with EU AI transparency requirements while safeguarding consumer trust and commercial performance.
Denne afhandling undersøger, hvordan forskellige måder at oplyse brugen af AI‑genererede billeder påvirker forbrugernes reaktioner i mode-e-handel i lyset af EU’s AI-forordning, som kræver tydelig mærkning af AI-indhold. Tidligere forskning har primært fokuseret på, om AI bliver oplyst eller ej, mens denne undersøgelse ser på, hvordan tre konkrete oplysningsformater – statisk, progressiv og begrundet – påvirker tillid, oplevet brandeffort og købsintention. Med afsæt i Persuasion Knowledge Model og Cognitive Load Theory gennemføres en kvalitativ, eksplorativ undersøgelse baseret på 18 semistrukturerede interviews med europæiske forbrugere, der alle aktivt handler tøj online. Deltagerne eksponeres for én af tre videoreklamer med forskellige AI-oplysningsformater, mens visuel elicitation bruges til at indfange deres kognitive og følelsesmæssige reaktioner. Resultaterne viser, at formatet for AI-oplysning betyder mere end blot tilstedeværelsen af en oplysning. Statisk oplysning skaber øjeblikkelig, men moderat skepsis, der i høj grad afhænger af tidligere kendskab til brandet. Progressiv oplysning udløser de mest negative reaktioner, fordi den forsinkede afsløring ofte tolkes som bevidst vildledning. Den begrundede oplysning, hvor AI præsenteres som et menneskestyret hjælpemiddel, reducerer derimod tendensen til algoritmeaversion, øger oplevet brandeffort og hjælper med at bevare købsintentionen for de fleste deltagere. Teoretisk bidrager studiet til Persuasion Knowledge Model ved at vise, at forbrugeres opmærksomhed på påvirkningsforsøg ikke kun aktiveres af selve oplysningssignalet, men også afhænger af timing, framing og kontekstuel begrundelse. Praktisk giver afhandlingen mode-e-handelsvirksomheder konkrete retningslinjer for, hvordan de kan leve op til EU AI-forordningens gennemsigtighedskrav og samtidig beskytte forbrugertillid og kommercielle resultater.
[This abstract has been generated with the help of AI directly from the project full text]