Scholarlink

<p>Advances in artificial intelligence (AI) have facilitated the development of AI-generated instructors and synthetic voices for use in educational videos. While existing research often categorizes instructors as either real or AI-generated and voices as either human or synthetic, this binary classification may not fully capture the complexity of learners’ perceptions in multimedia learning. This study identifies perceived human-likeness as a key psychological construct and examines how different combinations of instructor type (real vs. AI-generated) and voice type (human vs. synthetic) influence learners’ perceived human-likeness, motivation, and knowledge retention. A 2 × 2 between-subjects experimental design was adopted, with 130 participants randomly assigned to 1 of 4 conditions: (1) real instructor with human voice, (2) real instructor with synthetic voice, (3) AI-generated instructor with human voice, or (4) AI-generated instructor with synthetic voice. The results showed no direct effects of either instructor or voice type on motivation and knowledge retention. However, mediation analysis revealed a significant indirect pathway: AI-generated instructors were perceived as less human-like than real instructors, which reduced learners’ motivation and, in turn, undermined knowledge retention. These findings have practical implications for educators and instructional designers. While AI-generated instructors offer scalable and cost-effective options for video production, their perceived human-likeness should be carefully considered to support learners’ motivation and cognitive outcomes.</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

Real versus AI-generated instructors and human versus synthetic voices in educational videos: Impacts on perceived human-likeness, motivation, and knowledge retention

  • 期刊论文

  • 发表日期：

    2026年7月3日

查看全文
  • Ruiqi Deng,
  • Ziluo Zhang,
  • Xinlu Yu,
  • Maoli Jiang,
  • Yilin Wang

摘要

Advances in artificial intelligence (AI) have facilitated the development of AI-generated instructors and synthetic voices for use in educational videos. While existing research often categorizes instructors as either real or AI-generated and voices as either human or synthetic, this binary classification may not fully capture the complexity of learners’ perceptions in multimedia learning. This study identifies perceived human-likeness as a key psychological construct and examines how different combinations of instructor type (real vs. AI-generated) and voice type (human vs. synthetic) influence learners’ perceived human-likeness, motivation, and knowledge retention. A 2 × 2 between-subjects experimental design was adopted, with 130 participants randomly assigned to 1 of 4 conditions: (1) real instructor with human voice, (2) real instructor with synthetic voice, (3) AI-generated instructor with human voice, or (4) AI-generated instructor with synthetic voice. The results showed no direct effects of either instructor or voice type on motivation and knowledge retention. However, mediation analysis revealed a significant indirect pathway: AI-generated instructors were perceived as less human-like than real instructors, which reduced learners’ motivation and, in turn, undermined knowledge retention. These findings have practical implications for educators and instructional designers. While AI-generated instructors offer scalable and cost-effective options for video production, their perceived human-likeness should be carefully considered to support learners’ motivation and cognitive outcomes.