Generative AI in Media Communication: A Critical Review of Themes, Ethics, and Research Gaps

Abstract

The recent, unparalleled upsurge in generative AI technologies—from groundbreaking advancements like GPT-4, DALL-E 2, and Midjourney—has swiftly integrated them into professional media and communication domains, including journalism, public relations, and advertising. While some recent scholarship documents this significant transformation, the relevant information remains notably dispersed across computer science, ethics, and communication studies, underscoring a pressing need for cohesive synthesis. This paper, therefore, applies a rigorous systematic literature review, involving an extensive search of Web of Science, Scopus, ACM Digital Library, and Google Scholar, to identify peer-reviewed papers published from 2020 to 2024. The final corpus of 68 articles was thematically analyzed, revealing three principal research areas: (1) AI's complex interplay with automated and augmented media production workflows; (2) an emergent typology of ethical crises and dilemmas, often shadowed by epistemic, labour, and relational concerns; and (3) preliminary efforts at formulating normative guidance for the ethical and responsible production and dissemination of AI algorithms. Finally, this research delineates theoretical and methodological limitations, most notably the disproportionately low number of audience reception studies and a preponderance of research focusing on audiences in the Global North. To address these lacunae, we outline a comprehensive agenda for further research, designed to foster new theoretical models for non-human (or algorithmic) communication, explore socio-economic impacts on media workers, and investigate implications for democratic discourse.