The influence of social media deepfake images on political ideology and polarization: the mediating roles of cognitive load and confirmation bias

As political deepfakes become increasingly common on social media, understanding their impact on political attitudes and polarization is essential. This study investigates how exposure to political deepfake images influences perceived political ideology and political polarization, with cognitive load and confirmation bias acting as mediating variables. Drawing on data from 490 social media users in Pakistan, Structural Equation Modelling (SEM) was used to examine the relationships between deepfake exposure, psychological processing, and political outcomes. The results reveal that exposure to deepfake images significantly increases both cognitive load and confirmation bias, which in turn shape perceived political ideology. This perceived ideology then strongly predicts political polarization. All direct and indirect relationships were statistically significant. Grounded in Visual Literacy Theory, the findings highlight how manipulated visual content affects political perceptions through cognitive mechanisms. This study offers theoretical contributions to the media effects literature and practical implications for promoting visual literacy and reducing polarization in digital political environments.