ARTIFICIAL INTELLIGENCE AS FEEDBACK GENERATOR IN ACADEMIC WRITING INSTRUCTION

Abstract. The integration of AI-powered tools for providing feedback on academic writing is considered an innovative approach in university language education. This method requires teachers to write structured texts, known as prompts, which can be interpreted and understood by generative models of artificial intelligence. These prompts are being developed by many Western universities. This research is based on AI prompts created by professors of British and American universities to provide AI-generated feedback on students’ academic texts. The prompts were specifically designed for Chat GPT-4, which is a large language model. This study involved a content analysis of these prompts to identify common characteristics in their communication content and to examine how these characteristics relate to the ideas expressed by their authors. The analysis revealed that the authors of the prompts seem to be aware of generative artificial intelligence’s limited capabilities to process students’ academic texts. They view artificial intelligence as a tool allowing students to prepare multiple drafts of their texts before submitting the final version for teacher assessment. Despite the modern research trend advocating for a model that emphasizes student engagement with various forms of feedback, the content of the prompts still reflects a teacher-centered feedback model.