Visual evidence still king in digital era? The Impact of Instagram Knowledge, Age, and Source Credibility on the Detection of deep-fake AI-Generated Images on Instagram.

Abstract

The increasing reliance on social media platforms such as Instagram for news consumption, coupled with advances in AI that enable the production of hyper-realistic deep-fake images, necessitates an investigation into the public's susceptibility to such manipulated media. Therefore, this study examined how age, application-specific knowledge of Instagram, and source credibility influence people's ability to detect AI-generated deepfake images on Instagram. The between-subjects experiment included 88 participants and manipulated source credibility (legitimate and credible vs. unknown and not credible). Both conditions presented participants with identical sets of authentic and deep-fake images, created by AI-tool Midjourney. Participants were then asked to evaluate the perceived realism of each image, with responses ranging from "definitely not real" to "real". Results confirmed the hypothesis that older participants were more likely to misidentify deepfakes as authentic. However, this effect did not apply to authentic images. Contrary to expectations, Instagram knowledge and source credibility did not significantly impact deep-fake detection. The findings underscore age as a key factor in detecting Instagram deepfakes, however furt.