Abstract With rapid advances in Artificial Intelligence (AI) over the last decade, schools have increasingly employed innovative tools, intelligent applications and methods that are changing the education system with the aim of improving both user experience and learning gain in the classrooms. Even though the use of AI to education is not new, it has not unleashed its full potential yet. Much of the available research looks at educational robotics and at non-intelligent robots in education. Only recently, research has sought to assess the potential of Socially Assistive Robots (SARs), including humanoids, within the domain of classroom learning, particularly in relation to learning languages. Yet, the use of this form of AI in the field of mathematics and science constitutes a notable gap in this field. This study aims to critically review the research on the use of SARs in the pre-tertiary classroom teaching of mathematics and science. Further aim is to identify the benefits and disadvantages of such technology. Databases' search conducted between January and April 2018 yielded twenty-one studies meeting the set inclusion criteria for our systematic review. Findings were grouped into four major categories synthesising current evidence of the contribution of SARs in pre-tertiary education: learning gain, user experience, attitude, and usability of SARs within classroom settings. Overall, the use of SARs in pre-tertiary education is promising, but studies focussing on mathematics and science are significantly under-represented. Further evidence is also required around SARs' specific contributions to learning more broadly, as well as enabling/impeding factors, such as SAR's personalisation and appearance, or the role of families and ethical considerations. Finally, SARs potential to enhance accessibility and inclusivity of multi-cultural pre-tertiary classroom is almost unexplored.

中文翻译：

---

学前教育中社会辅助机器人的系统评价

摘要 随着人工智能 (AI) 在过去十年中的快速发展，学校越来越多地采用创新工具、智能应用程序和方法来改变教育系统，旨在改善课堂中的用户体验和学习收益。尽管人工智能在教育中的应用并不新鲜，但它还没有充分发挥其潜力。许多可用的研究着眼于教育机器人和教育中的非智能机器人。直到最近，研究才试图评估社交辅助机器人 (SAR)，包括类人机器人，在课堂学习领域，尤其是在学习语言方面的潜力。然而，这种形式的人工智能在数学和科学领域的使用构成了该领域的一个显着差距。本研究旨在批判性地回顾在高等数学和科学课堂教学中使用 SAR 的研究。进一步的目标是确定这种技术的优点和缺点。2018 年 1 月至 4 月期间进行的数据库搜索产生了 21 项符合我们系统评价设定的纳入标准的研究。调查结果分为四个主要类别，综合了 SAR 在高等教育中的贡献的当前证据：学习增益、用户体验、态度和 SAR 在课堂环境中的可用性。总体而言，在高等教育中使用 SAR 是有希望的，但专注于数学和科学的研究的代表性明显不足。还需要关于 SAR 对更广泛学习的具体贡献的进一步证据，以及促成/阻碍因素，例如 SAR 的个性化和外表，或家庭的作用和道德考虑。最后，特别行政区在提高多元文化前高等教育课堂的可及性和包容性方面的潜力几乎未被探索。