With the rapid development of science and technology, the traditional mode of teaching is inefficient and difficult to flexibly respond to the needs of knowledge updating, and generating content and applications based on AI has become an important way to solve this problem. According to the form of interaction in the digital exhibition hall, the article proposes SinGAN model and uses the multi-head self-attention mechanism to coordinate the overall features and detailed features in the generated adversarial network image, and to deal with the large range of dependencies in the image. The proposed AI-generated content and SinGAN image processing method are applied in the teaching of practical courses using the course “Digital Electronics Technology and Application” of a university in Guangdong Province, which specializes in electronic information and engineering, as an experimental object. The experiment shows that the percentage of content with a content quality score of 0.6 to 1.0 reaches 75.7%. As the course progresses, the keyword coverage rate reaches 0.996, and AI-generated content is efficiently applied in the course. The student performance of the experimental class with AI-generated content and image processing method teaching mode and the regular class with traditional teaching mode were 80.75 and 67.91 respectively, and the sample t-test for the significance of the student performance of the two classes was P=0.006, which showed a significant difference in the students’ performance between the two teaching modes. Students’ satisfaction with the new teaching mode is high, indicating that the AI-generated content and image processing methods proposed in the article have been well applied in education reform.