How journalism researchers navigate the AI hype: research orientations and intervention recommendations

As AI becomes increasingly integrated into news production, research on its implications for journalism abounds. However, how these studies normatively judge and provide interventions, which is crucial for journalism researchers and practitioners to progress in the discipline, largely remains unexplored. To fill this gap, extending the dichotomy between administrative and critical research orientations proposed by Lazarsfeld [(1941). Remarks on administrative and critical communications research. Zeitschrift für Sozialforschung 9 (1941): 2–16. https://doi.org/10.5840/zfs1941912], this study examines how researchers interpret their findings normatively and propose interventions beyond academia. We develop a framework categorizing journalism research orientations into four types: administrative, critical, advocacy, and problem-solving, while incorporating the comparative media systems theory to contextualize the reviewed studies. Analyzing 144 empirical articles from the Web of Science database on AI’s role in editorial decision-making and news dissemination, we find a prevalent negativity bias in interpreting results, with approximately half of the studies offering intervention recommendations. Critical research dominates, highlighting negative outcomes without constructive solutions. Conversely, one-third of the studies adopt a problem-solving orientation, addressing challenges while proposing actionable interventions. These recommendations target individual news workers, intra-organizational dynamics, and inter-organizational relationships. While both liberal and democratic corporatist systems have a similar proportion of critical and problem-solving research, little attention is given to the ethical, political, and legal tensions between journalists and external forces such as platforms, governments, and markets. We urge future research to engage with broader regulatory and societal contexts, providing evidence-based interventions to support the professionalism of journalism in the evolving digital public sphere shaped by AI. © 2025 The Author(s).