From Content Creators to Output Managers: Journalist-LLM Interaction and Its Consequences for Job Satisfaction and Professional Identity
Abstract
The role LLMs play in journalism introduces new and underexplored questions. Prior research has mainly examined journalists’ use of AI tools through surveys and interviews, but offers limited insights into concrete journalist-LLM interactions and how these influence the profession. In this paper, we present findings from ten contextual interviews with journalists across four media companies. We identify five consequences arising from journalist-LLM interaction: reduced human contact, a move from content creators to output managers, changing skill requirements, prolonged interaction without productive outcome, and experiences of frustration and fear. These consequences may influence professional identity and job satisfaction among journalists. We also find that some journalists perceive LLMs in ways that can foster overtrust and a false sense of shared responsibility. The use of LLMs in journalism appears highly individualized, highlighting a need for common ground and best practices. We argue that HCI can help address this need by deepening understanding of these interactions and designing LLM-based tools aligned with journalistic values and workflows.