When seeing is not believing: self-efficacy and cynicism in the era of intelligent media

Abstract The rise of deep synthesis technology has garnered significant attention with the advancement of artificial intelligence (AI). Social media platforms are increasingly populated with news content featuring AI-generated images. When browsing fragmented online information, it remains uncertain whether users pay attention to the source of this content. This study examines AI self-efficacy and user cynicism, exploring users’ cognitive and perceptual attitudes, as well as their behavioural tendencies toward news images generated by deepfake technology. The results of two experiments (N = 1826) suggest that when AI self-efficacy is lowered, users display greater cynicism toward the authenticity of news. Furthermore, differences in information processing strategies were observed depending on the content features of the news. Specifically, users with reduced AI self-efficacy were more likely to express cynicism toward high-relevance news compared to low-relevance news. Similarly, they were more likely to question the authenticity of low-risk news compared to high-risk news.