How Can We Fight Partisan Biases in the COVID-19 Pandemic? AI Source Labels on Fact-checking Messages Reduce Motivated Reasoning

Upon a surge of misinformation surrounding COVID-19, fact-checking has received much attention as a tool to fight the rampant misinformation. However, such correction efforts have faced challenges from partisans' biased information processing. For example, partisans trust or distrust a fact-checking message based on whether the message benefits or harms their supporting party. To minimize such politically biased processing of corrective health information, this experimental study examined how different source labels of fact-checkers (human experts vs. AI vs. user consensus) affect partisans' perceived credibility of fact-checking messages about COVID-19. Our findings showed that AI and user consensus (vs. human experts) source labels on fact-checking messages significantly reduced partisan-based motivated reasoning in evaluating fact-checking message credibility.