Combating deepfakes: the impact of warning label features on user engagement and verification intention of deepfake videos

Advances in artificial intelligence have led to the rise of deepfake technology, which poses significant challenges for information credibility due to its ability to create hyper-realistic, fabricated media content. In response, social media platforms have implemented warning labels to mitigate the impact of misleading content, including deepfakes. However, there is limited research on how specific warning label features impact users’ responses to deepfake content. Informed by construal level theory and the literature on colour psychology, we conducted a between-subjects experimental study with 317 social media users in China, to explore how content concreteness and colour of warning labels influence user engagement with deepfake content and their intention to verify it. The results showed that content concreteness of warning labels decreased user engagement with deepfake content, and increased their verification intention. Moreover, concreteness and colour interacted to influence verification intention. Specifically, red warning labels led to higher verification intention when paired with high-concreteness content, while this effect was not observed with white labels. These findings contribute to deepfake research and offer practical insights for designing effective warning labels on social media platforms to combat misinformation.