This research aims to contribute to the unexplored field of audience studies with a focus on the credibility of automated journalism. In this paper, we take a systematic look into the perceived credibility of robot-written news articles, searching specifically for differences and similarities between journalists and news consumers. In total, 232 native Dutch speakers (the language of the experiment) took part in this research, and among them were 64 journalists. The participants were asked to evaluate the perceived levels of the expertise and trustworthiness of four news articles based on algorithms outlined in the data-to-speech system (D2S) and created by Theune et al. (2001). We used a 2 (author: computer or journalist) x 2 (story topic: sport or finance) between-subject design to determine the perceived credibility of the news writer (source) and the contents of the news story (message).
Within the group of news consumers, no main effect was found. News consumers perceived the levels of the trustworthiness and expertise of the computer writer and journalist equally. Within the group of journalists, we found a significant effect on the perceived trustworthiness of the news source. In our experiment, journalists perceived the trustworthiness of a journalist to be much higher than that of the computer. Further, journalists perceived the expertise of the computer to be higher than the news consumers perceived it to be. Finally, the story topic has an influence on the item’s perceived trustworthiness. Overall, respondents perceived the trustworthiness of a sports article to be lower than that of a finance article.
It will be interesting to investigate this topic further, as it is possible that these differences between journalists and consumers will increase along with the rise in automated storytelling.