INSON–SUN’IY INTELLEKT KOMMUNIKATSIYASINI LINGVISTIK TAHLIL QILISHNING ASOSIY KATEGORIYALARI VA MEZONLARI
Litsenziya: cc-by
Annotatsiya
The article substantiates the main categories of linguistic analysis of human–artificial intelligence communication and the corresponding evaluation criteria. It shows that the asymmetry of this communication, the absence of paralinguistic means, the emergence of the prompt as a distinct communicative genre and the limitation of the system’s memory to the context window require a reinterpretation of the units of analysis. The categories are grouped into five sets – pragmatic, discursive, cognitive, epistemic and stylistic-social – and for each set a system of criteria with indicators and a three-level rating scale is proposed. Illustrative Uzbek examples support the need to rate a single response separately on each criterion and show that the priority of criteria depends on the purpose of communication. A four-stage rating procedure is proposed to make the analysis reproducible and verifiable. The results can be applied to the study of AI communication in Uzbek, in philological education and in the linguistic evaluation of such systems.
Hali tarjima qilinmagan