Assessing the Performance of Large Language Models for Bilingual Term Extraction in Interpreting and Translation.

Abstract

Large Language Models provide us with efficient tools for various Natural Language Processing tasks. This research evaluates four AI tools—ChatGPT, DeepSeek, Copilot, Gemini and Manus—for extracting Spanish-Arabic bilingual terminology for translation and interpreting purposes. Given the lack of research on non-Indo-European languages, the study assesses each tool using Precision, Recall, F-score, and Accuracy to determine their suitability for professional use and improve AI-driven terminology management.