Abstract

This study addresses the factors that increase the persuasive impact of testimonial narrative messages on artificial intelligence (AI). In particular, the effect on two variables that, to date, have not been explored in this field is analyzed: the attitudes toward AI (positive versus ambivalent) expressed by the protagonist of the narrative message (a testimonial in audiovisual format) and the role of participants’ prior beliefs about AI. An online experiment (N = 652) was carried out to contrast the effect of attitudinal similarity on identification with the protagonist of the narrative message and the indirect effect on attitudes and intention to use AI. The results showed that the message whose protagonist expressed positive attitudes toward AI induced greater identification only in those participants with previous positive beliefs. In contrast, the message whose protagonist expressed ambivalent attitudes toward AI induced greater identification only among participants with previous negative beliefs. In addition, identification and cognitive elaboration were found to mediate the effect of attitudinal similarity on the attitude toward and intention to use AI. These findings are discussed in the context of narrative persuasion research and the development of campaigns for improving social perceptions of data science.