The use of artificial intelligence-assisted teaching technology for the teaching reform of higher vocational English courses Civic and political education can not only enhance the topicality and ideology of higher vocational English teaching but also promote the development of teaching in the direction of intelligence and precision. The study aims to extract Civics and Politics elements from the higher vocational English curriculum using natural language processing. The BERT pre-training model is used to obtain the word vectors. Then sentences that can be used as text summaries are selected using cosine similarity calculation. Finally, the selected sentences are fine-tuned by the RoBERTa-Seq2Seq model to remove some repetitive words and obtain text summaries. On the basis of the RoBERTa-Seq2Seq model, an enhanced word vector feature dimension, pointer generation network, and coverage penalty mechanism are added to solve related problems in the process of model application. After the curriculum integration based on the model of this paper, the core literacy competence of higher vocational students has significantly improved. The students’ English scores increased by 23.5 points compared to those before curriculum integration. It shows that the model can effectively help teachers improve the effectiveness and quality of classroom education.