The Susceptibility of Humanity to AI Visual Content: Do Age and Exposure Play a Role?

Ashlynn Perez

Co-Presenters: Individual Presentation

College: College of Liberal Arts

Major: BA.PSYCHOLOGY

Faculty Research Mentor: Jane O'Brien

Abstract:

As artificial intelligence (AI)-generated content has become more realistic and prevalent, distinguishing between AI-generated images and humanmade images has become more challenging. The present study examined whether age and prior experience with AI influence individuals’ ability to discern between AI images and human images. Using a within-subjects design, 60 participants ages 18 years and older were shown 18 images (nine AI-generated and nine humanmade) and rated whether the images were AI or human using a five-point scale. Results indicated that both age and prior experience with AI were associated with detection accuracy. Younger participants and those with prior experience generally performed better. Despite this, across all groups, detection of humanmade images was lower than AI-generated images. These findings suggest that AI-generated images have reached a level of believability that not only complicated the identification of AI content, but also leads to confusion about the authenticity of media people consume. The results align with prior research on the perceived realism of AI-generated media and add to existing literature by demonstrating the role of age and experience in AI detection. Overall, the findings highlight the importance of digital literacy and the evaluation of visual media in an increasingly AI-driven environment.