Algorithmic Bias or Algorithmic Reconstruction? A Comparative Analysis Between AI News and Human News

Despite a substantial body of scholarship at the intersection of artificial intelligence (AI) and journalism, it remains relatively unexplored as to how AI-generated news is different from news produced by professional journalists in terms of news bias. To fill the gap, this study compares human versus GPT-2-generated news in terms of the linguistic features, tone, and bias toward gender and race/ethnicity on two highly controversial issues, namely abortion and immigration, using news transcripts from CNN and Fox News. In doing so, the study adopts a mixed-method content analysis approach, including dictionary and coreference analysis, topic modeling and semantic network analysis, and manual content analysis. The results reveal that although AI news differs from human news in terms of language features and thematic areas, machine news is not necessarily more biased compared to human news regarding gender and race/ethnicity. Implications are discussed for future scholarship on algorithmic bias in lieu of the roles that AI-generated news may play in journalism and democracy.