А COMPARATIVE ANALYSIS OF FIRST-PERSON PRONOUN USE IN AI AND HUMAN ESSAYS

This study comparatively analyzed the use of the personal pronoun “I” in AI-generated and human-written essays through an AntConc analysis of two self-made corpora of texts, to (dis)prove the hypothesis: AI-generated essays will use the first-person pronoun "I" significantly less frequently than human-written essays. The first corpus of essays in this study, all generated by ChatGPT, consisted of 100 essays totaling 29,442 words. The second corpus of 100 essays, written by humans, included texts with a total word count of 39,878. The frequency analysis revealed a significant disparity in the use of the pronoun “I” between the two corpora. In the AI-generated essays, "I" was used only 15 times, contrasting sharply with the human-written essays, where “I” appeared a total of 523 times. These findings showed the noteworthy differences in the narrative styles of AI and human writers, showing the potential limitations of free Generative AI models in replicating human-like self-reference and subjectivity in written discourse.