CCEF
CCEF 2025 主旨演讲1 回顾—— Professor Rupert: 教育该如何回应生成式人工智能的挑战? How should education respond to the challenge of generative AI?发表时间：2025-08-22 13:04
一、AI时代的教育之问 Wegerif教授首先提出，教育是关于设计的——不是描述世界，而是改变世界。那么在很多任务AI比人类做得更好的情况下，我们该教什么？怎么教？为什么要教？长久以来，技术被认为是教育的一种附加或补充，但是教授认为技术其实是教育的核心。
接下来Wegerif教授对比了两类实证研究，一类表明频繁使用AI工具会损害批判性思维能力，而另一类研究则证实AI工具有潜力促进用户的深度思维。所以问题不在于AI本身，而在于教育系统和教学设计。学生想要拿高分，所以倾向使用AI替自己完成任务，结果导致思维退化（不是自己不聪明，是教学鼓励了这种行为）。教师不能简单归咎于工具，而应反思教学设计和评估机制。
学生可以将AI视为一起探索可能性的思维伙伴，而非答案的提供者，这里Wegerif教授建议了探索式提问（Exploratory inquiry）、有目标的优化（Targeted refinement）、生成式协作（Generative engagement）三种交互方式。教育中不应仅关注具体工具的使用技巧，而应为学生创造持续探索与反思的空间，培养学生与AI共同面对复杂问题的能力。
接下来，Wegerif教授举出了由澳大利亚科技大学（UTS）开发的AI对话工具Qreframer和美国德州的一个AI赋能的学校Alpha Schools的例子，说明AI有成为认知工具和重塑教学结构的可能。 二、AI支持的对话教育
Wegerif教授指出，对话是教育的目的，也是教育的途径。回顾历史会发现，追溯到苏格拉底，思考需要通过对话——一种囊括不同视角、不断追问和解释的方式——来实现。同样，中国的对话教育也由来已久，老子、孔子等先哲在经典中多次提及。
对话意味着将两个独立的意识，移动到一个共享的对话和思维空间。而AI可以作为这个过程的“旁观者”甚至“引导者”，当学习者带着自己的问题和主观能动性使用AI时，AI及其背后的人类知识才能被激活和产生意义，实现人机协同智能（co-intelligence）。
以上观点在他的书籍《The Theory of Educational Technology: Towards a dialogic foundation for design》（中文版名称为《教育技术的对话之道》），以及《Rethinking Education Theory: Education as Expanding Dialogue》有更详细的阐述。 供稿人 于玻，剑桥大学教育系博士生，北京师范大学互联网教育硕士与教育技术学学士，研究方向包括对话教育、人工智能的教育应用、学习科学与学习分析等。 审核：楼菲凡 Professor Rupert Wegerif from the Faculty of Education, University of Cambridge, who is also the founder and academic lead of the Digital Education Futures Initiative (DEFI), delivered a keynote speech titled《教育该如何回应生成式人工智能的挑战（How should education respond to the challenge of generative AI?）》 1. The Question of Education in the Age of AI Professor Wegerif began by asserting that education is aboutdesign—not merely describing the world, but changing it. In a context where AI can outperform humans in many tasks, what should we teach, how should we teach, and why should we teach? For a long time, technology has been regarded as an add-on or supplement to education, but Professor Wegerif argued that technology is, in fact, central to education. He then contrasted two types of empirical studies: one showing that the frequent use of AI tools can harm critical thinking skills, and another demonstrating that AI tools have the potential to foster deep thinking. The issue, therefore, does not lie with AI itself, but with the education system and instructional design. Since students are motivated to achieve high grades, they tend to use AI to complete tasks on their behalf, which leads to a decline in thinking skills—not because they lack intelligence, but because teaching practices encourage such behaviour. Teachers should not simply blame the tool, but instead reflect on their pedagogical design and assessment mechanisms. Students can treat AI as a thinking partner for exploring possibilities rather than a provider of answers. In this regard, Professor Wegerif proposed three modes of interaction:Exploratory inquiry,Targeted refinement, andGenerative engagement. Education should not merely focus on technical skills for using specific tools, but instead create spaces for sustained exploration and reflection, fostering students’ capacity to work with AI in tackling complex problems. Finally, Professor Wegerif presented two examples to illustrate AI’s potential as a cognitive tool and as a catalyst for reshaping pedagogical structures: the AI dialogue toolQreframerdeveloped by the University of Technology Sydney (UTS), andAlpha Schoolsin Texas, USA, an AI-enabled educational institution. 2. AI-Supported Dialogic Education Professor Wegerif emphasised that dialogue is both the goal and the means of education. Looking back through history, one finds that as far back as Socrates, thinking was understood to require dialogue—a process that incorporates multiple perspectives, continual questioning, and explanation. Similarly, traditions of dialogic education have long existed in China, with ancient thinkers such as Laozi and Confucius frequently referring to it in the classics. Dialogue entails bringing two independent consciousnesses into a shared space of conversation and thought. AI can act as an “observer” or even a “facilitator” in this process. When learners engage with AI from their own questions and with active agency, the AI—together with the human knowledge behind it—can be activated to generate meaning, thus enablingco-intelligencebetween humans and machines. These ideas are elaborated in greater detail in his booksThe Theory of Educational Technology: Towards a Dialogic Foundation for Design(Chinese edition titled《教育技术的对话之道》) andRethinking Education Theory: Education as Expanding Dialogue. Editor: Yu BoBo Yu is a PhD student in Education, Faculty of Education, University of Cambridge. MA in Internet Education and BSc in Educational Technology, Faculty of Education, Beijing Normal University. Research Area: Dialogic Education, AI in Education, the Learning Sciences and Learning Analytics. Reviewer:Feifan Lou - 关于我们About Us - 小红书Rednote：CCEF剑桥中国教育论坛 LinkedIn&Instagram：Cambridge China Education Forum Twitter：Cambridge China Education Forum (@Forum Cambridge) 官网 Official Websie：https://www.cambridgecef.net/ 剑桥大学官方学生会注册信息Registered Student Society at University of Cambridge： https://www.cambridgesu.co.uk/organisation/6531/ |
