arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

日本学术振兴会科学研究费补助金(KAKENHI)资助项目摘要中生成式AI潜在应用的分析

Analysis of Potential Generative AI Use in Abstracts of KAKENHI-Funded Projects

Hitoshi Koshiba, Miki Ida-Kimura

arXiv 2608.20674首次发表:更新:

AI 中文总结

本研究针对KAKENHI科学研究(C)类项目摘要,分析2022-2026财年AI生成摘要的占比变化,发现2025年起占比上升,2026年约20%,生成式AI已在该类项目涉及的多领域普及。

AI 中文摘要

本研究分析了日本学术振兴会科学研究费补助金(KAKENHI)科学研究(C)类资助项目的摘要被归类为AI生成的程度。分析覆盖了2022财年至2026财年五年间资助的项目,被归类为AI生成的摘要数量从2025财年开始增长,2026财年约20%的摘要被归类为AI生成。尽管比例存在一定差异,但在多个评审类别中均观察到被归类为AI生成的摘要,这些发现表明生成式AI的使用已开始在科学研究(C)所代表的众多研究领域中普及。

英文摘要

This study analyzes the extent to which abstracts of projects funded under the Scientific Research (C) category of the Grants-in-Aid for Scientific Research (KAKENHI) were classified as AI-generated. The analysis covers projects funded over the five-year period from FY2022 to FY2026. The number of abstracts classified as AI-generated began to increase in FY2025, and approximately 20% of the abstracts were classified as AI-generated in FY2026. Although the proportions varied to some extent, abstracts classified as AI-generated were observed in many review categories. These findings indicate that the use of generative AI has begun to spread across many research fields represented in Scientific Research (C).

Comments14 pages, 7 figures, Earlier Japanese version submitted to Jxiv

论文原文

arXiv 摘要页 · PDF 原文 · HTML 原文

↑