Artificial intelligence has caused changes in higher education, and traditional ideological and political discourse has become difficult to handle. Many situations are: existing methods are not flexible enough when teaching conditions change, and have weak connections with cultural values. Words may be formally correct, yet seem divorced from students’ actual experiences, and are limited to the role of strengthening cultural identity. This study develops an adaptive method to improve the university’s ideological and political discourse system. Combining data analysis and dynamic adjustment, this method enables discourse strategies to respond to changing educational environments without departing from cultural guidelines. Instead of fixing discourse design, it takes the continuous coordination of educational goals, student participation and value transmission as the core.The empirical results on different datasets are relatively stable. On the HE ideology course dataset, the proposed method achieves an accuracy of 90.23% and an F1 score of 89.09%, which is 1.56 percentage points higher than the strongest baseline.On the AI-augmented political discourse dataset, the accuracy rate is 91.89%, the F1 score is 90.94%, which is 1.55 percentage points higher than the best comparative model. Overall, these findings indicate that AI-based methods can provide assistance for improving ideological and political education in higher education, and can also provide practical support for strengthening cultural identity amid rapid technological changes.
