O POPRAWNOŚCI JĘZYKOWEJ TEKSTÓW GENEROWANYCH PRZEZ SI NA PRZYKŁADZIE CHATUGPT.

The article aims to assess the linguistic accuracy of texts generated by ChatGPT, based on queries that mirror the instructions found in the 2023 secondary school final exam in Polish language and literature. The analysis focused on identifying and classifying language errors and evaluating their frequency. The results of the analysis revealed a varied level of accuracy in AI-generated texts. Most errors occurred in texts containing more complex constructions, where correct formulation required a comprehensive understanding of grammatical rules. The conclusions drawn from the analysis offer insights into the limitations of AI, which users should be mindful of when editing texts using chatbots. [ABSTRACT FROM AUTHOR]