Artificial intelligence from a linguistic perspective - opportunities, functions and challenges in natural language processing

Artificial intelligence (AI) represents one of the most significant technological achievements of the twenty-first century. Its rapid development has created new opportunities while simultaneously raising numerous linguistic, communicative, and interdisciplinary questions. This paper examines the role of systems based on Large Language Models (LLMs) in natural language processing, with particular emphasis on their ability to analyze, interpret, and generate linguistic structures. The theoretical section explores the relationship between linguistic form, meaning, and communicative context, as well as the extent to which these systems can develop competencies comparable to human linguistic and pragmatic competence. Particular attention is devoted to the capabilities of contemporary models, such as ChatGPT, Gemini, and Copilot, in processing syntactic, semantic, pragmatic, and discourse-related aspects of language. Furthermore, the paper proposes a systematic framework for classifying their functions based on two criteria: (1) linguistic-communicative function and (2) the relationship to knowledge and information sources. Within the first criterion, four principal categories are identified: text transformation, content generation, analysis and categorization of linguistic data, and interaction and assistance through intelligent dialogue systems. The analysis demonstrates that contemporary AI tools are capable of producing coherent and grammatically accurate texts and support a wide range of language-related applications, including machine translation, text summarization, paraphrasing, and information extraction. Nevertheless, fundamental challenges remain concerning genuine semantic understanding, implicit meanings, speech acts, culturally conditioned patterns of communication, and context-sensitive interpretation. The paper emphasizes the need for a critical evaluation of both the capabilities and the limitations of artificial intelligence, as well as its application in contemporary linguistic research, education, and academic communication.
Keywords: artificial intelligence, Large Language Models (LLMs), natural language processing, pragmatic competence, linguistic-communicative functions, academic communication.