Scholarlink

Personalized generative AI (GAI) holds promise in supporting student creativity, though evidence regarding its effectiveness remains limited. This study employed a within-subject design to compare the effects of standard versus personalized GAI on students’ creative performance, self-efficacy, and attention allocation. Each of the 62 students engaged in two creativity tasks: one with standard GAI (ChatGPT-4) providing generic feedback, and the other with a personalized GAI tutor (Mr. Ranedeer), implemented via prompt engineering to deliver goal-aligned feedback tailored to students’ needs. Results indicated that personalized GAI improved creative performance in output quality and elaboration, boosted self-efficacy, and increased gaze transitions across task-relevant areas of interest (AOIs) as evidence of attention allocation. Furthermore, higher self-efficacy was associated with more transitions across these areas. These findings contribute to understanding the mechanisms through which personalized GAI influences student creativity, providing valuable insights into the design of strategies that promote GAI-supported creative development.

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

Effects of Personalized Generative AI on Student Creativity: Evidence from Creative Performance, Self-efficacy, and Attention Allocation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Yixuan Chen,
  • Xuemei Wu,
  • Yihua Zhong,
  • Changqin Huang

摘要

Personalized generative AI (GAI) holds promise in supporting student creativity, though evidence regarding its effectiveness remains limited. This study employed a within-subject design to compare the effects of standard versus personalized GAI on students’ creative performance, self-efficacy, and attention allocation. Each of the 62 students engaged in two creativity tasks: one with standard GAI (ChatGPT-4) providing generic feedback, and the other with a personalized GAI tutor (Mr. Ranedeer), implemented via prompt engineering to deliver goal-aligned feedback tailored to students’ needs. Results indicated that personalized GAI improved creative performance in output quality and elaboration, boosted self-efficacy, and increased gaze transitions across task-relevant areas of interest (AOIs) as evidence of attention allocation. Furthermore, higher self-efficacy was associated with more transitions across these areas. These findings contribute to understanding the mechanisms through which personalized GAI influences student creativity, providing valuable insights into the design of strategies that promote GAI-supported creative development.