大语言模型在辅助外语口语教学中的应用困境:语言数据的局限性探讨  

Challenges in Applying Large Language Models to Foreign Language Speaking Instruction:A Discussion on the Limitations of Linguistic Data

作　　者：丛慧君 Cong Huijun(Harbin Normal University,Harbin,Heilongjiang,150025)

机构地区：[1]哈尔滨师范大学,黑龙江哈尔滨150025

出　　处：《现代英语》2024年第24期79-81,共3页Modern English

摘　　要：随着大语言模型(ChatGPT,Gemini)的快速发展,其在外语口语教学中的应用潜力日益凸显。这些模型通过自动生成语言内容和提供即时反馈,能够有效增加学生的语言输入量和互动频次,成为传统教学模式的有力补充。尽管大语言模型在口语教学中带来了许多创新机遇,但仍面临诸多挑战,如在语言真实性和示范性等方面存在不足以及可能会出现不可预料的政治偏差。文章探讨了大语言模型在外语口语教学中的应用前景与挑战,并提出了针对性的改进建议,旨在为大语言模型赋能外语教学模式发展创新提供有益的参考。With the rapid advancement of large language models(ChatGPT,Gemini),their potential in foreign language speaking instruction has become increasingly prominent.By generating language content and providing instant feedback,these models effectively enhance students'language input and interaction frequency,serving as a valuable supplement to traditional teaching methods.Despite their innovative potential,large language models still face challenges,particularly in terms of linguistic authenticity,instructional reliability,and the risk of unforeseen political bias.This paper explores the opportunities and challenges of applying large language models in foreign language speaking instruction and proposes targeted improvements to support their integration into language education.

关 键 词：大语言模型 外语口语教学 语言真实性 语言示范性 

分 类 号：H319.3[语言文字—英语]