A DISCOURSE ANALYSIS OF THE LINGUISTIC PATTERNS OF AI CHATBOTS

Abstract

Artificial Intelligence has become a transformative factor in shaping mediated interaction, altering how humans construct meaning and manage communication with chatbot users. This study aimed to analyse the linguistic patterns of AI chatbots, particularly ChatGPT, through a discourse analytical framework. Using a thematic synthesis framework of fifteen peer-reviewed studies published between 2020 and 2025, the research employed a qualitative meta-synthesis approach to analyse recurrent linguistic, and socio-pragmatic features of AI-generated discourse. The reviewed studies indicated that chatbots demonstrate advanced lexical diversity, syntactic complexity, and formal cohesion but lack spontaneity, emotional depth, and contextual adaptability. At the pragmatic level, AI discourse relied on politeness markers and neutral expressions that created an impression of professionalism with limited interpersonal warmth and authenticity. Although rhetorical strategies such as logical appeals appeared frequently, they were interpreted in relation to how communicative intent and interpersonal alignment were negotiated within the discourse rather than as purely literary devices. The findings suggest that while AI chatbots emulate human linguistic structures, their discourse remains constrained by superficial pragmatics and formulaic coherence. These insights contribute to understanding how computational language production reflects communicative competence rather than mere linguistic competence, emphasizing the importance of sociolinguistic awareness in model design. Therefore, developers should prioritize pragmatic adaptability, context sensitivity, and grammatical accuracy to enhance the naturalness and authenticity of human-AI communication.