INGLIZ VA O‘ZBEK TILLARIDA SUN’IY INTELLEKT DISKURSINING LINGVOPRAGMATIK XUSUSIYATLARI

Maqolada ingliz va o‘zbek tillarida yaratilgan sun’iy intellekt (SI) diskursi — katta til modellari asosidagi chat-botlar javoblari — lingvopragmatik jihatdan qiyosiy tahlil qilinadi. Tadqiqot korpusini ChatGPT, Gemini va Claude tizimlaridan bir xil so‘rovlar asosida olingan ingliz va o‘zbek tilidagi javoblar tashkil etadi. Nutq aktlari nazariyasi, Grays kooperativlik tamoyili va Braun–Levinsonning xushmuomalalik nazariyasi asosida SI nutqida assertiv aktlarning ustunligi, direktivlarning yumshatilgani, xedjlash vositalarining yuqori chastotasi, o‘zini ifodalash (self-reference) va diskursiv markerlarning shablonlashgan qo‘llanishi aniqlandi. O‘zbek tilidagi SI matnlarida ingliz tilidan kalkalangan pragmatik formulalar ("Ajoyib savol!", "Umid qilamanki, bu yordam beradi") va o‘zbek muloqot madaniyatiga xos hurmat shakllari o‘rtasidagi nomuvofiqlik kuzatildi. Natijalar tarjimashunoslik, chet tili o‘qitish va o‘zbek tilidagi SI tizimlarini takomillashtirish uchun ahamiyatga ega.

The article presents a contrastive linguopragmatic analysis of artificial intelligence (AI) discourse, understood as the responses of chatbots based on large language models, in English and Uzbek. The corpus consists of responses generated by ChatGPT, Gemini and Claude to identical prompts in both languages. Drawing on speech act theory, Grice’s Cooperative Principle and Brown and Levinson’s politeness theory, the study identifies the predominance of assertives, mitigated directives, a high frequency of hedging devices, specific patterns of self-reference and formulaic discourse markers. Uzbek AI texts display a mismatch between pragmatic formulae calqued from English ("Ajoyib savol!", "Umid qilamanki, bu yordam beradi") and the norms of Uzbek communicative culture. The findings are relevant for translation studies, foreign language teaching and the development of Uzbek-language AI systems.