Exploring Correction Format and Literacy Intervention as Boundary Conditions of Bot-Based Fake News Corrections

Artificial intelligence (AI) offers a promising approach to governing fake news on social media, yet its effectiveness relative to human fact-checkers and the conditions shaping user acceptance remain unclear. Drawing on the Computers Are Social Actors paradigm and the Elaboration Likelihood Model, we conducted two large-scale online surveys with Chinese netizens (combined N = 3,846). Study 1 (N = 2,146) used a 2 (corrector: human vs. bot) × 2 (correction: simple tag vs. detailed rebuttal) design across 17 fake news stories. Bot-based corrections were generally as effective as human corrections; however, humans were most effective with simple tags, whereas bots required detailed rebuttals to achieve stronger outcomes. Study 2 (N = 1,700) tested news literacy and algorithm literacy interventions. Results provided partial evidence that news literacy messages improved correction effectiveness. These findings support combining AI-based corrections with critical thinking interventions to improve fake news governance.