The impact of diagnosis and information processing bias created by ai generation: from algorithm recommendations to the intention to discern misinformation of Vietnamese users

Abstract

Misinformation has become a major problem for digital platforms and social networks in the age of content consumption governed by algorithms, especially with the quick advancement of generative artificial intelligence (GenAI). This study aims to investigate how algorithmic features and cognitive factors affect users' intentions to recognize misinformation about GenAI. In particular, the study looks at how users' information evaluation behaviors are influenced by perceived accuracy, perceived credibility, and diagnosticity. Utilizing survey data gathered from Vietnamese users, a quantitative research approach was used to meet these goals. The study assessed the connections between algorithmic characteristics, cognitive dimensions, and misinformation discernment goals using statistical analytic methods and structural equation modeling (SEM). The findings demonstrate that algorithmic components significantly influence diagnosticity, which enhances customers' perceptions of accuracy and credibility. These views have a favorable effect on users' desire and ability to identify misinformation. The results also reinforce the need of AI literacy as a "cognitive shield" that improves users' ability to evaluate information systems critically in the digital world. By expanding knowledge of misinformation discernment behavior in AI-driven contexts, the work makes a theoretical and practical contribution. Furthermore, it has managerial implications for enhancing digital literacy and creating tactics to assist users in protecting themselves against the expanding "information pandemic" in the era of artificial intelligence.