Artificial Intelligence (AI) has been widely used in the social and legal fields, and ChatGPT, after AI painting, has once again set off a wave of discussion on whether AI and its generated works can obtain legal protection. Starting from the theoretical orientation of the origin of ChatGPT legal governance, this paper proposes the legal positioning and layered governance framework of ChatGPT application. It explores the role mechanism of ChatGPT empowering legal modernization and combs through the realistic dilemmas of ChatGPT-generated content data compliance legalization. To effectively analyze legal risks in the process of the ChatGPT application, data crawling technology and SMOTE oversampling technology are utilized to obtain ChatGPT application data and produce datasets. The Stacking integration strategy is introduced to combine the Random Forest in the Decision Tree Algorithm, GBDT algorithm, and Support Vector Machine to construct the legal risk prediction model of the ChatGPT application. For the effectiveness of the model, the ChatGPT application dataset is used to analyze the accuracy, ROC curve, and AUC value, which provides a reference for improving the legal system related to the ChatGPT application. The results show that the accuracy of the SVM classifier reaches 0.839, the correctness of the GBDT model is 0.947, and the AUC value of ChatGPT legal risk prediction based on the Stacking integration strategy is 0.947. Based on the inspiration of the decision tree algorithm, the improvement of the legal system related to the ChatGPT application should be improved in terms of generating content and allocating risk. Based on the insights of the decision tree algorithm, the improvement of the legal system related to the ChatGPT application should be made based on the dimensions of generation content and risk allocation.