CHATGPT DAN GEMINI: ANALISIS STILISTIK SYAIR KAAB BIN ZUHAIR

Kajian ini meneliti keupayaan dua model kecerdasan buatan generatif, iaitu ChatGPT dan Gemini, dalam menganalisis ciri stilistik syair Arab klasik melalui tiga bait terpilih daripada Banat Su‘ad karya Kaab bin Zuhair. Analisis dijalankan berdasarkan kerangka stilistik Leech dan Short yang merangkumi empat aspek utama, iaitu leksikal, sintaksis, retorik serta kohesi dan konteks. Kajian ini menggunakan pendekatan kualitatif perbandingan dengan pemilihan tiga bait secara purposif bagi mewakili tiga bahagian utama qasidah, iaitu nasib, deskripsi dan pujian. Bagi memastikan perbandingan yang lebih setara, kedua-dua model diberikan arahan analisis yang sama berdasarkan kategori stilistik yang telah ditetapkan. Output yang dihasilkan kemudiannya dibandingkan bagi mengenal pasti persamaan, perbezaan, kelebihan dan keterbatasan setiap model. Dapatan menunjukkan bahawa kedua-dua model mampu mengenal pasti unsur stilistik utama dalam syair yang dianalisis, namun memperlihatkan kecenderungan analisis yang berbeza. ChatGPT menunjukkan kekuatan dalam analisis teknikal bahasa, khususnya penghuraian struktur sintaksis, fungsi unsur dalam ayat dan penggunaan istilah retorik yang lebih tepat. Gemini pula lebih menonjol dalam penghuraian makna leksikal, nilai estetika serta pengaitan unsur bahasa dengan konteks budaya dan sejarah puisi. Dalam beberapa keadaan, Gemini menghasilkan analisis retorik dan kontekstual yang lebih luas, manakala ChatGPT memberikan penghuraian linguistik yang lebih sistematik dan terperinci. Perbezaan ini menunjukkan bahawa kedua-dua model berpotensi saling melengkapi dalam menghasilkan analisis stilistik yang lebih komprehensif. Walau bagaimanapun, penggunaan model AI dalam analisis sastera Arab klasik masih memerlukan penilaian dan semakan manusia, khususnya dalam menentukan ketepatan istilah balaghah, memahami konteks budaya dan sejarah serta menilai kesesuaian interpretasi terhadap teks asal. This study examines the capabilities of two generative artificial intelligence models, ChatGPT and Gemini, in analysing the stylistic features of classical Arabic poetry through three selected verses from Kaab bin Zuhair’s Banat Su‘ad. The analysis is based on Leech and Short’s stylistic framework, which comprises four main dimensions: lexical features, syntactic features, rhetorical features and cohesion and context. A qualitative comparative approach was employed, with three verses purposively selected to represent the principal sections of the qasidah, namely nasib, description and praise. To ensure a more balanced comparison, both AI models were given identical analytical prompts based on the predetermined stylistic categories. Their outputs were subsequently compared to identify similarities, differences, strengths and limitations in their analyses. The findings indicate that both models were able to identify major stylistic elements in the selected verses, although they showed different analytical tendencies. ChatGPT showed greater strength in technical linguistic analysis, particularly in explaining syntactic structures, identifying the functions of linguistic components and employing rhetorical terminology with greater precision. Gemini, on the other hand, performed more strongly in interpreting lexical meanings, aesthetic values, and the relationship between linguistic elements and the cultural and historical contexts of the poem. In several instances, Gemini provided broader rhetorical and contextual interpretations, while ChatGPT offered more systematic and detailed linguistic explanations. These differences suggest that the two models may complement each other in producing a more comprehensive stylistic analysis. Nevertheless, the use of AI models in analysing classical Arabic literature continues to require human evaluation and verification, particularly in determining the accuracy of balaghah terminology, understanding cultural and historical contexts and assessing the appropriateness of interpretations in relation to the original text.