2026 ASEE Annual Conference & Exposition

This is a evidence-based practice full paper.
There are rapid progresses and extensive application of artificial intelligence in almost all kinds of industries in recent decades. In 2017, Next Generation Artificial Intelligence Development Plan was Issued by China's State Council and various forms of activities to popularize AI were encouraged and supported in the plan. Especially, AI education programs were proposed including AI-related courses in elementary and secondary schools. Plenty of devices, robots, integrated development kits with series of curriculums and books were developed and supplied by companies to various training classes inside and outside schools.
However, it's an urgent need to build a standard curriculum and evaluation methodology for the AI education of elementary and secondary school students. Based on the program of "Definition and Selection of Competencies: Theoretical and Conceptual Foundations" (DeSeCo) proposed by OECD, the core competences and values for Chinese students’ development proposed by Beijing Normal University, and curriculum standard of information science and technology proposed by Ministry of Education of China, an new evaluation model for AI core competences was built by us with four dimensions: 1. common AI knowledge and skills; 2. high level thinking about AI application; 3. cooperative innovation of human-machine systems; 4. ethics and social responsibility in AI era.
Furthermore, a practical framework was proposed base on the evaluation model of AI education. There were mainly two key characters of the framework: a). 1 + X, which means the evaluation should be done from one common part, AI fundamentals and several different subjects, such as scratch, python programming, micro:bit or Arduino based application etc. b). step-by-step, which means every evaluation subject includes several steps, 1st-8th levels generally, and the evaluation was also divided into two phases, popularization and deep application. In conclusion, the evaluation framework had many different subjects with multiple dimensions.
The AI core competence evaluation (AICE) program has been carried out once a quarter since November of 2021. More than 75 thousands students from elementary and secondary schools had participated the program, and many of these students have past 6 levels of a subject. Three study cases were illustrated in this paper. The changes of every student who learning AI courses and past multiple levels were shown in the first case, and the differences of plenty of students in one evaluation were described in the second case. Finally, the improvement of AI competences of 20 high school students was evaluated before and after a smart car course in a semester in the third case.