A COMPARATIVE STUDY OF GEN-AI AND HUMAN ANALYSIS OF EVALUATIVE MEANING IN MEDIA REPORTS ON HERDSMEN ATTACKS IN NIGERIA

Semantic analysis of texts, including news reports, involves various analytical approaches, such as examining evaluative language. Since journalists play a significant role in shaping public perceptions and understanding of events, the language used in news reporting deserves critical attention. In Nigeria, one recurring issue in newspaper coverage is the conflict between herders and farmers, particularly attacks on farmers and farmlands. Reports of these incidents are often presented from differing, and sometimes overlapping, perspectives. Recent advances in Generative Artificial Intelligence (Gen-AI) have demonstrated their potential for language analysis, including interpretation and evaluation of texts. To examine this potential, the present study compares Gen-AI and human analyses of evaluative meaning in media reports on attacks of herdsmen in Nigeria. The study aims to assess how both Gen-AI chatbots and human analysts identify and interpret evaluative language and the stances conveyed in selected newspaper reports. The dataset comprised ten reports on attacks on herdsmen published in The Guardian and The Punch during the first half of 2021. These reports were analysed using the Appraisal Theory framework. The findings show that both Gen-AI and human analysts were effective in identifying evaluative meaning, although their analytical outputs differed. Gen-AI successfully detected recurring evaluative patterns and produced consistent analyses but was less effective in capturing nuanced meanings shaped by contextual, cultural, and ideological factors. Human analysis, by contrast, demonstrated greater contextual sensitivity, critical reasoning, and nuanced interpretation, reflecting a deeper understanding of Nigeria’s socio-cultural realities. However, human interpretations were found to be more subjective and less systematic than those produced by Gen-AI. The study concludes that while Gen-AI is a powerful tool for textual analysis, human judgment remains essential for interpreting evaluative meaning. Therefore, a collaborative approach in which artificial intelligence complements, rather than replaces, human expertise is recommended to achieve the most robust analytical outcomes.