Trust and Deception in AI-Generated Advertising: The Double-Edged Role of AI Realism, Disclosure, and Consumer AI Literacy

Abstract Generative artificial intelligence (AI) is blurring the boundary between human-made and machine-generated communication. As AI-generated content becomes increasingly realistic, audiences must evaluate not only what a message communicates but also how and by whom it was produced. This study examines how perceived AI realism shapes trust and deception judgments in AI-generated advertising and whether these effects depend on AI disclosure and consumers' AI literacy. A between-subjects experiment was conducted with 224 adult consumers. AI realism (high vs. low) and AI disclosure (disclosed vs. undisclosed) were manipulated in a 2 x 2 experimental design, while AI literacy was treated as a measured continuous moderator. Participants viewed an AI-generated advertisement for a fictitious dietary supplement brand and evaluated advertising trust, brand trust, and perceived deception. Conditional process analyses using PROCESS Macro Model 3 were conducted with gender and age as control variables. The findings reveal a double-edged role of AI realism. Higher AI realism increased advertising trust and brand trust across experimental conditions. Disclosure strengthened the trust-enhancing effect of AI realism, while AI literacy further shaped this effect only for brand trust. For perceived deception, higher AI realism increased deception judgments most strongly when AI use was not disclosed. When disclosure was provided, conditional probing showed that this deception-enhancing effect weakened and was not significant at higher levels of AI literacy, although the three-way interaction was not statistically significant. These results show that AI realism is not merely a positive cue of content quality. Rather, its effects depend on transparency practices and consumers' interpretive competence regarding AI-generated content.