Reconstruction and Empirical Research on Full-Process Maritime English Teaching in Higher Vocational Colleges Supported by Dual Collaborative Generative AI Models

Abstract

Maritime English teaching in higher vocational colleges is plagued by practical dilemmas including insufficient provision of occupation-based scenarios, limited oral training channels, unimplemented student stratification, and rigid summative assessment systems. To advance the digital transformation of maritime specialized English education, this study adopts situated learning, task-based language teaching, and multi-dimensional process evaluation theories to construct a three-stage full-process teaching framework featuring the collaborative operation of DeepSeek and Doubao. The framework forms a closed-loop teaching system of "pre-class differentiated guided learning – in-class situated interactive training – post-class data-driven assessment". This research aims to develop a low-cost, replicable digital teaching implementation scheme, providing a localized practical paradigm for the intelligent teaching reform of English for Specific Purposes (ESP) in higher vocational education.