摘要：

【目的】 测试ChatGPT在中文科技期刊摘要文字编辑中的实用性,为科技期刊编辑人员适应人工智能发展趋势提供参考。【方法】 以50篇中文科技期刊摘要为测试文本,采用文本分析法,对比ChatGPT 3.5编辑前后文本的差异。【结果】 ChatGPT可以有效校正语法错误,简化语言表述并规范学术用语,但存在提供错误信息、错用专业术语等缺陷。【结论】 ChatGPT能够辅助编辑人员进行文字编辑工作,但编辑的人工判断仍不可替代。我国应重视培育科技期刊编辑人员的智能化信息素养,同时积极研发我国自主的辅助编辑人工智能工具。

关键词: ChatGPT, 人工智能, 科技期刊, 文字编辑

Abstract:

[Purposes] This study aims to explore the applicability of ChatGPT in the text editing for abstracts in Chinese scientific journals and provide a reference for editors to adapt to the development of artificial intelligence. [Methods] A total of 50 abstracts in Chinese scientific journals were taken as the test texts, and the differences before and after editing by ChatGPT 3.5 were analyzed. [Findings] ChatGPT corrects the grammar errors, polishes the language, and standardizes academic expression. However, it has defects such as providing error message or misusing specialized terms. [Conclusions] ChatGPT could assist editors in text editing, while the manual judgment is still irreplaceable. We should give priority to the cultivation of intelligent information literacy and the research & development of artificial intelligence tools for editors in China.

Key words: ChatGPT, Artificial intelligence, Scientific journal, Text editing