THE USE OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES FOR ANALYZING MARITIME ACCIDENT REPORTS IN TEACHING MARITIME ENGLISH TO ENGINE CADETS

Abstract

The article substantiates the pedagogical value of using artificial intelligence (AI) technologies to work with authentic maritime accident reports in Maritime English training for engine cadets. Accident investigation reports are viewed as a rich source of specialized vocabulary, communication breakdown patterns, and causal relations that support safety culture. The paper analyses didactic barriers to using such documents in class (large volume, high readability difficulty, and legal register) and explains how natural language processing (NLP) tools and large language models can facilitate adaptation without losing technical accuracy. A staged AI-supported model is proposed: (1) selecting relevant report fragments; (2) semantic tagging of terminology and causality markers; (3) prompt design for generating summaries, glossaries, and B1–B2 learning tasks; (4) validation through comparison with the original report and expert review. Learning activities are exemplified in engine room communication (reporting a failure to the chief engineer, issuing urgency messages, and completing technical forms) and in analyzing human communicative error. Safe classroom use is outlined through retrieval-augmented generation (RAG) and mandatory fact-checking practices to minimize hallucination risks. The approach reduces teacher preparation time, supports individualization, and increases learner motivation while maintaining CLIL coherence. Expected learning outcomes include expansion of cadets’ active technical lexicon, improving comprehension of complex reports, and higher accuracy in concise incident reporting aligned with IMO Standard Marine Communication Phrases. The model also develops critical digital literacy by training learners to evaluate AI outputs against verified evidence in real settings.

Received: November 18, 2025
Accepted: December 15, 2025
Published: January 16, 2026