How many university students are using GenAI in their assessments? A scoping review.

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

New generative artificial intelligence (GenAI) tools like ChatGPT show high levels of performance on many assessment methods used commonly in higher education, potentially raising a serious challenge to assessment validity. However, it is unclear whether students are actually using GenAI in their assessments. Media stories focus on individual survey studies which report high use levels, but these stories do not reflect an analysis of all the data on student GenAI use, or how quickly this is changing and what students are using GenAI for. This paper describes a scoping review wherein the scale and speed of student GenAI use in assessments is quantified by analysing data from all identified survey studies conducted since the launch of ChatGPT. The results suggest that over 80% of students are now using GenAI in their assessments, although the number who report using it to complete entire assessments remains low. Assessment use tracks general use, suggesting that GenAI is simply becoming a part of everyday life. Most studies used small, non-representative samples, again highlighting a need for more rigorous sampling methodology in higher education research.The extraordinary speed with which GenAI tools have entered use in assessment in higher education has urgent implications for what, how and why we assess.

Links and sources

Comments