粤港澳大湾区智能教育需求测算研究

Guangdong-Hong Kong-Macao Greater Bay Area (G-H-M GBA), as the economic, technological and educational highland of China, has unique user preferences and technical features in its demand for Artificial Intelligence education (i.e. AI technology taking part in the process of education). On the basis of accurately defining the major five types of applied forms of AI education in G-H-M GBA, this project builds a supply-demand index system consisting of the demand index, usage index and potential value; it collects around 20,000 sample data of the eight major groups of population in the eleven cities in G-H-M GBA; and it comprehensively examines the picture of AI education demand in G-H-M GBA – according to which, we compare the structures of the eleven cities’ index scores, so as to provide a quantitative reference for the intensive development of the AI education within the same city circle of G-H-M GBA in the future. Furthermore, this project creatively transforms the user portrayal technology that is based on Model-base CF, and it describes the high index scores, namely, the portrait of the group that has a high demand of AI education – in order to provide quantifiable, displayable reference and indication for the decision making in the development of AI education, the development of relative technologies, and the investment in relevant fields in G-H-M GBA, as well as for the educational cooperation and joint development of the three areas, two systems and three language families in G-H-M GBA.

粤港澳大湾区作为中国的经济高地、科技高地与教育高地，在智能教育（指人工智能技术参与教育过程）的需求上，有着独特的用户偏好与技术特征。本项目在精准界定湾区智能教育五大类应用形态的基础上，构建由需求指数、使用指数、潜力值组成的供需指标体系，收集湾区11城市8大人群2万余样本数据，全面摸底湾区智能教育的需求图景——由此比对11城市的指数得分结构，为湾区未来同城圈智能教育的集约化发展提供量化参考。进一步，创新性改造“基于模型的协同过滤算法（Model-base CF）”用户画像技术，描画高指数得分、即智能教育高需求人群的画像——为湾区智能教育发展决策、相关技术开发决策、相关领域投资决策，以及湾区范围内三个地区、两种制度、三种语系之间的教育协作与融合发展，提供可量化的、可显示化的参考与启示。

从项目背景来看，从2020年初到2022年末，即本项目整体执行期间，国家施行较为严格的疫情防控政策，我国教育系统在疫情防控期间面向全国亿万学生开展的大规模在线教育是历史上绝无仅有的，在全球也尚属首次，展现了我国本土数字教育特别是高阶水平的智能教育的巨大潜力，用户偏好、技术特征等供需规律，以及借助技术进行跨地域教育协作与融合发展的路径与前景。从重要结果和关键数据来看，本项目克服疫情期间实地调研全面受限的约束，通过在线调研、调整数据采集和研究方式等，在一定程度上突破原申报书粤港澳大湾区范围内的抽样方案，收集了华东、华中、华北、东北、东南、西南、西北七大区域31个省级行政区的“停课不停学”实践数据（填答教师人数共计近18万，填答家长人数共计180余万），形成并发表了一系列学术成果。从科学意义来看，上述成果系统总结了我国数字教育和智能教育在2020年“停课不停学”行动中的本土规律和技术潜力，并与全球实践进行了国际比较研究，比较全面地反映了这一全球最大规模的信息化教学、智能化教学的社会实验全景，也比较全面地显现了教育系统在线抗疫的中国经验，并进一步对当前我国“人工智能+教育”的发展进程进行梳理总结，包括发展现状总结、供需规律、短板和挑战总结，对未来发展前景提出重点建议。