Registered report: Does repetition increase the credibility of AI-generated images?

Abstract

In an era where digital imagery has become ubiquitous, the ease of creating fake or AI-generated images poses a critical challenge to credibility assessment. Repetition increases the credibility of written and verbal information. Does this credibility gain also apply to pictures? Participants were exposed to both genuine and AI-generated images, followed by a credibility assessment phase including repeated and new images. We found that participants judged repeated images as more credible than new images, regardless of whether those images were genuine or AI-generated. This repetition effect remained consistent across diverse image categories with low between-image variability. These findings highlight a potent mechanism for the spread of visual misinformation: Images repeatedly shared across social media platforms may gain unwarranted credibility through mere exposure, regardless of their authenticity. However, the precise cognitive mechanisms underlying the repetition effect for pictures-and whether they differ from those established for verbal information-remain important questions for future research. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

Ähnliche Arbeiten