ChatGPT人机对话式翻译研究

打开文本图片集

关键词：ChatGPT模型；人机对话；神经机器翻译；启示

中图分类号：H319.3 文献标识码：A 文章编号：1009-895X（2025）02-0118-10

DOI： 10.13256/j.cnki.jusst.sse.230918456

AStudy on ChatGPT's Human-MachineInteractive Translation Performance

LI Mei，KONG Delu （School ofForeign Languages， Tongji University， Shanghai 20oo92， China）

Abstract： The current research explores ChatGPT's human-machine interactive translation performance in machine translation in comparison with neural machine translation （NMT） systems represented by Google Translate in English-Chinese translation. Case analyses show that ChatGPT overrides NMT systems in that it is able to provide multiple translations， produce translations of various styles，and adjust translations based on the meta information of the source text.In contrast， Google Translate does a beter job than ChatGPT in terms of accuracy， adequacy， and consistency. In short， ChatGPT’ s human-machine translation performance presents implications for translators， students， and software developers.

Keywords： ChatGPT；human-machine interaction； neural machine translation； implications

近年来，翻译界对于技术的冲击已不再陌生，诸多“机器翻译替代人工翻译”的话题已是老生常谈[1-3]。（剩余15359字）