Initiating a novel elementary school artificial intelligence-related image recognition curricula

The international community is accelerating the establishment of an AI education curriculum system, however, there is a vacuum in the primary education level. We have started to develop a novel AI-related education framework for elementary students. Firstly, the current AI-related curricula were reviewed; Secondly, four kinds of guidelines for the construction of AI curriculum systems in the elementary educational stage are proposed to standardize subsequent AI-related teaching activities; Thirdly, a five-level model including perceptualization, digitization, conceptualization, algorithmization, and specialization was set up and the corresponding AI teaching curricula according to the five-step teaching model were developed and implemented; Finally, a deep learning algorithm of convolutional neural network for image recognition was introduced to the 5th-grade students in an elementary school. A positive evaluation of more than 80 points for the courses was given by the students, which confirmed the potentiality and feasibility of the current curriculum system for developing the intelligence of students.