On Their Own: Perception and Use of Generative AI among Ghanaian Broadcast Journalists

This study examines how Ghanaian broadcast journalists engage with generative artificial intelligence in their newsroom practice. Drawing on Shoemaker and Reese's sociology of journalism framework, the study conceptualizes generative AI as a sociologically mediated tool whose use is shaped by professional norms, newsroom routines, and organizational constraints rather than by technical affordances alone. Using an exploratory approach, the study analyzes data from a qualitative survey (n = 38) and semi—structured interviews (n = 4) conducted between January and April 2024. The findings show that while awareness and use of generative AI are widespread, journalists integrate these tools selectively into specific stages of news production, such as background research, story structuring, and proofreading. Journalists reported that their core editorial functions—including verification, authorship, and final publication decisions—remain under human control.This pattern reflects journalists' efforts to preserve normative boundaries around editorial authority and professional judgment. However, the absence of formal newsroom training and AI governance structures provided by news organizations limits journalists' preparedness, leaving many to rely on informal or externally sourced knowledge. This research presents both the opportunities and risks of generative AI use in Ghanaian broadcast journalism and underscores the need for newsroom-level training and policy development to support responsible and contextually grounded AI integration in the Global South.By foregrounding journalistic practice within news organizations, the study contributes to sociological understandings of AI integration, showing how journalists negotiate technological possibilities alongside professional expectations, newsroom routines, organizational conditions, and the expectations of actors within the broader social structure production.