Summary:

The development of artificial intelligence (AI) allows rapid creation of AI-synthesized images. In a pre-registered experiment, we examine how properties of AI-synthesized images influence belief in misinformation and memory for corrections. Realistic and probative (i.e., providing strong evidence) images predicted greater belief in false headlines. Additionally, we found preliminary evidence that paying attention to properties of images could selectively lower belief in false headlines. Our findings suggest that advances in photorealistic image generation will likely lead to greater susceptibility to misinformation, and that future interventions should consider shifting attention to images. 

Summary:

Corrections alone rarely erase the impact of misinformation. In this EEG study, we examine how the brain encodes and later evaluates corrected information. Providing an alternative explanation alongside a retraction significantly reduced the continued influence effect and enhanced neural markers of memory encoding and recollection during later truth judgments. The findings suggest that replacing misinformation with a plausible alternative strengthens corrective processing in memory, offering new insight into the neurocognitive mechanisms that make misinformation so persistent, and how to counter it more effectively.