This study examines the realization of illocutionary acts in ChatGPT’s responses to students, focusing specifically on representative and directive illocutionary acts based on John R. Searle’s Speech Act Theory (1979). As artificial intelligence becomes increasingly integrated into educational contexts, understanding how AI performs communicative actions through language is essential from a pragmatic perspective. This research employs a qualitative descriptive method to analyze 256 ChatGPT utterances collected from written interactions with 20 English Literature students. The students were asked to engage in natural conversations with ChatGPT, while the analysis concentrated exclusively on ChatGPT’s responses as the primary data. The findings reveal that representative illocutionary acts are dominant, occurring in 150 utterances (58.37%), while directive illocutionary acts appear in 106 utterances (41.63%). Representative acts are primarily used to explain, define, and evaluate information, positioning ChatGPT as an informative and explanatory interlocutor. Directive acts function mainly to provide guidance, suggestions, and encouragement, supporting students’ decision-making and problem-solving processes. This distribution indicates that ChatGPT balances informational and instructional roles in academic interaction. The study concludes that ChatGPT demonstrates emerging pragmatic competence in educational settings through its effective use of representative and directive illocutionary acts. However, the findings are limited to a specific academic context and two illocutionary categories. Future research is encouraged to explore other types of speech acts and broader educational environments to further examine AI-mediated academic discourse.