BADIIY TARJIMADA SUN’IY INTELLEKT: “PROMPT WRITING”NING TARJIMA BADIIYATINI QAYTA YARATISHDAGI AHAMIYATI

Ushbu ilmiy maqolada badiiy tarjimada sun’iy intellekt (SI) tizimlari, xususan, Yirik til modellari (LLM) va generativ platformalardan (ChatGPT, Microsoft Copilot) foydalanishning dolzarb muammolari filologik paradigma doirasida tadqiq etiladi. "Prompt-injiniring" (injenerlik buyruqlarini yozish) texnikasining badiiy matn estetikasini, milliy kolorit va lisoniy realiyalarni boshqa tilga o‘tkazish hamda tarjima badiiyatini qayta yaratishdagi ahamiyati chuqur tahlil qilingan. Abdulla Qodiriyning “O‘tkan kunlar” romani va uning uchta inglizcha tarjimasi (Carol Ermakova, Mark Reese, I.M.Tukhtasinov) empirik baza sifatida olinib, “Zero-shot”, “Few-shot” va “Chain-of-Thought” buyruq berish strategiyalarining samaradorligi qiyosiy o‘rganilgan. Tarjima natijalari ham metrologik (BLEU, METEOR), ham lisoniy-ekspertlik baholash usullari orqali tekshirilgan hamda SI yordamida tarjimadagi xatolarni aniqlash va tarjima transformatsiyalarini optimallashtirish metodologiyasi ishlab chiqilgan.

This scientific article examines the contemporary issues of integrating Artificial Intelligence (AI) systems, particularly Large Language Models (LLMs) and generative platforms (ChatGPT, Microsoft Copilot), into literary translation within the philological paradigm. The significance of "prompt engineering" (the art of crafting structured text commands) in reproducing the aesthetic quality, national coloring, and cultural realia of literary texts is thoroughly analyzed. Based on Abdulla Qodiriy's masterpiece novel “Days Gone By” (“O’tkan kunlar”) and its three distinct English translations (by Carol Ermakova, Mark Reese, and I.M.Tukhtasinov), the efficacy of “Zero-shot”, “Few-shot”, and “Chain-of-Thought” prompting methodologies is comparatively evaluated. The generated translations are assessed using both automatic metrological metrics (BLEU, METEOR) and qualitative human linguistic expertise. Furthermore, a methodology for error detection and translation transformation optimization via AI is established.
Keywords: literary translation, artificial intelligence, large language models (LLM), prompt engineering, Days Gone By, cultural realia, national coloring, BLEU, Chain-of-Thought, translation transformations.