Scholarlink

<p>The application of artificial intelligence (AI) is becoming increasingly prevalent in the education context. Many empirical studies have investigated the impact of AI on students’ writing performance. However, the results of these existing studies are inconsistent. At present, most reviews or meta-analyses focus on automatic writing assessment tools, while there are relatively few meta-analyses on the application of AI in writing. This study aims to perform a meta-analysis to determine the effect of AI on writing performance and to determine the factors that moderate this effect. The meta-analysis included 46 empirical studies published between January 2000 and September 2025. Pooled analysis via a random-effects model revealed that AI had a medium-sized effect on students’ writing performance, with Hedges’ <i>g</i> of 0.601 (95% CI [0.437, 0.766], <i>p</i> &lt; 0.001). Furthermore, this study examined the influence of moderating variables such as intervention duration, type of AI and way of AI application. The moderation analysis indicated that intervention duration and sample size affected AI writing performance significantly (Q_between = 19.141, <i>df</i> = 4, <i>p</i> &lt; 0.01; Q_between = 8.540, <i>df</i> = 3, <i>p</i> &lt; 0.05), whereas the effects of the other moderating variables were not statistically significant. However, some subgroup findings should be interpreted cautiously because certain categories included only a limited number of effect sizes. These findings indicate that AI technology has considerable potential to enhance students’ writing performance. This study offers suggestions regarding further research and the practical application of AI-assisted writing-related learning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The effect of artificial intelligence on students’ writing performance: a meta-analysis

  • 期刊论文

  • 发表日期：

    2026年7月17日

查看全文
  • Juan Wu,
  • Qiuxu Chen,
  • Huiting Jiang,
  • Xinyue Bai

摘要

The application of artificial intelligence (AI) is becoming increasingly prevalent in the education context. Many empirical studies have investigated the impact of AI on students’ writing performance. However, the results of these existing studies are inconsistent. At present, most reviews or meta-analyses focus on automatic writing assessment tools, while there are relatively few meta-analyses on the application of AI in writing. This study aims to perform a meta-analysis to determine the effect of AI on writing performance and to determine the factors that moderate this effect. The meta-analysis included 46 empirical studies published between January 2000 and September 2025. Pooled analysis via a random-effects model revealed that AI had a medium-sized effect on students’ writing performance, with Hedges’ g of 0.601 (95% CI [0.437, 0.766], p < 0.001). Furthermore, this study examined the influence of moderating variables such as intervention duration, type of AI and way of AI application. The moderation analysis indicated that intervention duration and sample size affected AI writing performance significantly (Q_between = 19.141, df = 4, p < 0.01; Q_between = 8.540, df = 3, p < 0.05), whereas the effects of the other moderating variables were not statistically significant. However, some subgroup findings should be interpreted cautiously because certain categories included only a limited number of effect sizes. These findings indicate that AI technology has considerable potential to enhance students’ writing performance. This study offers suggestions regarding further research and the practical application of AI-assisted writing-related learning.