Journalists on the Ethics of In-House AI Visuals: Not in the News and Never Photorealism

Manipulated news images have been present since the advent of photojournalism, yet the integration of AI-generated visuals within legacy media introduces novel ethical challenges, particularly concerning their capacity to undermine trust in visual journalism. While previous research has predominantly focused on AI in textual and editorial domains, this study addresses the ethical implications that arise as legacy media begins to adopt AI for in-house image production, and how journalists perceive that this may potentially shift the ontology of photojournalism. Drawing on semi-structured interviews with journalists from Swedish legacy media, the analysis is situated within the theoretical framework of applied journalism ethics, emphasizing credibility and accountability as foundational principles of professional standards in visual reporting from the perspective of mediated authenticity. The findings suggest that although Swedish outlets increasingly experiment with AI-generated imagery, they claim to refrain from deploying photorealistic visuals in news contexts. These reservations, shaped by professional visual literacy, center on transparency and truthfulness, highlighting broader concerns regarding authenticity and the preservation of public trust in journalistic practice.