Liwei Shen

This study investigates the efficacy of AI-generated visuals and potential credibility cues in correcting health-related misinformation. Using a pre-registered, factorial experiment ( N = 1,751), we tested two AI-generated visual enhancements (visual exemplars, infographics) and two credibility boosters (source tagging, partisan neutrality in posting history). Findings revealed small but significant advantages of visual exemplars, but not infographics, over text-only corrections in reducing misbeliefs, primarily through mitigating psychological reactance. However, credibility cues did not significantly enhance correction effectiveness. Overall, the results suggest that AI-generated visual exemplars offer incremental persuasive benefits beyond textual correction alone.

Following the end of the Cold War, the rise of democracy and globalization fostered optimism about global cooperation and economic integration but also sparked debates on equity, leading to challenges such as economic disparities, cultural displacement, and populist movements. Collective narcissism, a concept introduced to understand self-identity and politics amid perceived threats to status, describes beliefs that the in-group is exceptional but lacks external recognition and is associated with adverse outcomes. Drawing from the Social Identity Model of Deindividuation Effects, this study investigates collective narcissism expressions on social media, focusing on the “Stop the Steal” movement after the 2020 US Presidential Election. Analyzing 11,836 original posts and their conversation threads, we studied the nature of collective narcissistic expression and its contagion effects. Results highlight national identity as a central theme, showing that such expressions spread through increased user engagement and are contagious from original posts to replies.

In an era of intergroup tensions and conspiracy theories, our study examines whether threat-mitigation news frames can counter U.S. public beliefs in China-related conspiracy theories. Using a preregistered online survey experiment, we analyzed the effects of news articles that either emphasize or mitigate threats from China. Results show that threat-emphasizing frames did not heighten conspiracy beliefs, likely due to preexisting media framing. However, threat-mitigation frames effectively reduced conspiracy beliefs and improved factual understanding of China. This underscores the potential of responsible news reporting in fostering informed views, mitigating intergroup tensions.

Generative Artificial Intelligence (AI) models like DALL·E 2 can interpret prompts and generate high-quality images that exhibit human creativity. Though public enthusiasm is booming, systematic auditing of potential gender biases in AI-generated images remains scarce. We addressed this gap by examining the prevalence of two occupational gender biases (representational and presentational biases) in 15,300 DALL·E 2 images spanning 153 occupations. We assessed potential bias amplification by benchmarking against the 2021 U.S. census data and Google Images. Our findings reveal that DALL·E 2 underrepresents women in male-dominated fields while overrepresenting them in female-dominated occupations. Additionally, DALL·E 2 images tend to depict more women than men with smiles and downward-pitching heads, particularly in female-dominated (versus male-dominated) occupations. Our algorithm auditing study demonstrates more pronounced representational and presentational biases in DALL·E 2 compared to Google Images and calls for feminist interventions to curtail the potential impacts of such biased AI-generated images on the media ecology.