<p>Chatbots are making a strong entry into education, supporting both students and teachers. This study aims to deepen the understanding of the educational use of chatbots in science education, including their advantages and limitations. A scoping review of the articles published up to January 1, 2025, was conducted following PRISMA guidelines across the Web of Science, Scopus, and ERIC databases, using search terms related to science education and chatbots. From an initial pool of 608 articles, 40 met all inclusion criteria. Most of the selected studies were exploratory (32.5%), with fewer intervention-based designs. Chatbots in science education are promising tools but are still in their early stages, and this could explain the large number of exploratory studies found. ChatGPT is the most studied and has demonstrated excellent linguistic capabilities but needs to improve its scientific accuracy and analytical skills. Tutoring students is the most commonly found application of chatbots. They also have the potential to support teachers by reducing their workload, although empirical data are needed to confirm this. Integrating Artificial Intelligence literacy and critical thinking skills into curricula, alongside comprehensive teacher professional development, is crucial for the effective and responsible use of chatbots in science education.</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

Chatbots in Science Education: A Scoping Review of Early Empirical Evidence

  • Mario Calvo-Utrilla,
  • Esther Paños,
  • José-Reyes Ruíz-Gallardo

摘要

Chatbots are making a strong entry into education, supporting both students and teachers. This study aims to deepen the understanding of the educational use of chatbots in science education, including their advantages and limitations. A scoping review of the articles published up to January 1, 2025, was conducted following PRISMA guidelines across the Web of Science, Scopus, and ERIC databases, using search terms related to science education and chatbots. From an initial pool of 608 articles, 40 met all inclusion criteria. Most of the selected studies were exploratory (32.5%), with fewer intervention-based designs. Chatbots in science education are promising tools but are still in their early stages, and this could explain the large number of exploratory studies found. ChatGPT is the most studied and has demonstrated excellent linguistic capabilities but needs to improve its scientific accuracy and analytical skills. Tutoring students is the most commonly found application of chatbots. They also have the potential to support teachers by reducing their workload, although empirical data are needed to confirm this. Integrating Artificial Intelligence literacy and critical thinking skills into curricula, alongside comprehensive teacher professional development, is crucial for the effective and responsible use of chatbots in science education.