Despite growing newsroom interest in generative visual AI, the use of AI-generated images may backfire. An online experiment ( N = 1,171) shows that regardless of disclosure strategies, AI-generated images reduce trust in news organizations and acceptance of AI in journalism compared to real photos. However, attributing AI use to external motives (i.e., protecting victims’ privacy) elicits more favorable responses than internal attribution (i.e., saving newsroom costs) or simple disclaimers. Perceived legitimacy underlies audience response, and people’s predispositions toward AI moderated the effectiveness of disclosure strategies. These findings extend AI transparency research and help newsrooms to communicate AI use more ethically.
Li et al. (2026) studied this question.