Nationalism meets machine heuristics: Investigating the effect of AI’s “nationality” on the perceived credibility of AIGC

Against the backdrop of politicizing science and technology, it has nevertheless become a common practice to employ AI tools in political communication. This study delves into the tension between the supposedly high credibility of AI-generated content (AIGC) with its perceived objectivity and the potentially devastating impact of politicized tools. Specifically, we investigate the disparities in the perceived credibility of AIGC, stemming from various source types and nationalities. Using an online experiment conducted in China, we found that Chinese AI sources and human-AI hybrid sources with consistent Chinese national identity were perceived as more credible than their foreign (American) counterparts, whose effects were fully mediated by machine heuristics. In contrast, less credibility was directly attributed to human-AI hybrid sources with inconsistent human-AI nationalities (e.g., a Chinese journalist using American AI). Furthermore, although foreign sources were generally seen as less credible, the presence of AI in the source moderated this negative effect. These findings indicated users' perceptions of the political dimensions of AI technology.