Algorithmic revision: The role of generative AI and digital media in exam preparation among journalism students

Abstract

Generative artificial intelligence (GenAI) and digital media have rapidly become central to how students study, yet journalism students, whose profession is defined by verification, accuracy, and originality, remain under-researched, particularly in the high-stakes context of examination preparation and in settings outside the global North. This study describes the patterns, purposes, and perceived consequences of GenAI and digital-media use in exam revision among undergraduate journalism students, tests whether critical verification of AI output strengthens with academic seniority, and derives implications for curriculum design. A cross-sectional survey combining closed and open-ended items was administered to 65 undergraduate journalism students across four years of study during an examination period (April 2026). Closed items were analysed with descriptive statistics, cross-tabulation by year, and correlations; open-ended responses were analysed thematically using an inductive procedure. Adoption was widespread: 69.2% used these tools frequently or exclusively, led by ChatGPT (66.2%) and Google Gemini (35.4%). The dominant purpose was explaining complex concepts (mean = 3.92/5); 70.8% rated the tools very or highly effective and 61.5% reported reduced reliance on traditional materials. However, only 58.5% verified AI output consistently against authoritative sources, and, contrary to expectation, verification did not significantly increase with year of study. Qualitative themes were factual inaccuracy and "hallucination," divergence from lecturer-sanctioned content, and anxiety about eroded critical thinking; students’ outlook was markedly ambivalent, framing AI as a "double-edged sword." A value-action gap characterises journalism students’ AI use: uptake and perceived effectiveness are high while critical verification lags and does not mature with experience. Because students are already critically aware, programmes should teach AI verification and critical use as explicit, assessed professional competencies rather than relying on prohibition or on the assumption that judgement develops on its own.