ARTIFICIAL INTELLIGENCE AS A JUDGE OF STUDENT DEBATES (CASE STUDY OF THE BNTU SCIENTIFIC LIBRARY)

Abstract

This article analyzes the application of artificial intelligence as a judge in student debates. The relevance of the study stems from the need to find objective and technologically efficient assessment methods within the educational environment. The aim of this work is to analyze and evaluate the application of AI as a student debate judge, using the case study of the BNTU Scientific Library. The study involved developing a prompt for the AI judge based on the author's model of communication with AI and testing the Gemini 3.0 neural network as a judge in three debate sessions held between November 2025 and January 2026. The methodology included prompt engineering, multimodal analysis (handwritten text recognition), and a qualitative assessment of the AI's performance based on seven criteria. The results demonstrated high accuracy in text recognition, objectivity in assessment, detailed feedback, and the pedagogical value of applying AI. Key advantages (emotional neutrality, processing speed, gamification) and limitations (algorithmic bias toward academic style, infrastructure dependence, failure to account for non-verbal factors) were identified. The study demonstrates the technological readiness of Large Language Models (LLM) to perform assessment functions in educational activities and opens perspectives for the further integration of AI technologies into pedagogical practice.