Artificial Intelligence and Its Critical Capability: An Empirical Study on Arabic Poetry

Abstract

This research examines the potential of machines programmed for literary criticism by applying them to examples of Arabic poetry and comparing their performance with human cognition. The study concludes that AI models such as Claude and ChatGPT play a significant role in conducting comprehensive, high-speed critical analyses of texts. They are well-versed in both classical and modern critical methodologies; they elaborate on points a human critic might summarize, deconstruct vocabulary, explain meanings, and elucidate thematic unity, rhetorical imagery and its aesthetic function, poetic rhythm, and meters. However, these models exhibit certain limitations, such as occasional failures in the scansion of individual verses, a lack of understanding regarding the nuances of specific vocabulary, and potential errors in attributing authorship. Furthermore, their criticism is neither subjective nor grounded in aesthetic taste; instead, it relies on data previously stored by humans, as the AI ​​has not lived the poetic experience itself. Finally, total reliance on these tools in the linguistic realm poses a potential risk: the stagnation of human thought and a regression in intellectual progress.