Large-scale groundwater pollution risk assessment research based on artificial intelligence technology: A case study of Shenyang City in Northeast China

• Developed groundwater pollution risk assessment system suitable for large-scale ranges. • Integrated AI technology optimized the groundwater pollution risk assessment procedure. • Prediction performance of the CNN model is better than that of the traditional ML model. • Image segmentation technique was first applied to the intelligent partitioning of risk levels. Building on the concepts of groundwater vulnerability and pollution driving factors, this study incorporates the social response to environmental risks, selecting a total of 23 indicators to construct a groundwater pollution risk assessment system suitable for large regions. Furthermore, integrating artificial intelligence (AI) technology with the groundwater pollution risk assessment system optimized the evaluation process and, for the first time, attempted to apply image vision technology for risk level classification. Taking Shenyang City in China as an example, the developed groundwater pollution risk assessment system was applied and validated. The results show that the Convolutional Neural Network (CNN) model has the best prediction accuracy. Compared with the Artificial Neural Network (ANN) and Random Forest (RF) models, the performance evaluation parameters mean squared error (MSE), mean absolute error (MAE) and root mean squared error (RMSE) are closer to 0, and the coefficient of determination (R 2 ) is closer to 1. Unfortunately, the stability of the CNN model is poor. The coefficient of variation (CV) values of each evaluation parameter are higher than the CV values of the corresponding evaluation parameters in the ANN and RF models. The maximum CV value of the evaluation parameters can reach 16.23%. The groundwater pollution risk assessment results in Shenyang show that the area of high-risk areas accounts for 4.11%–5.23%, mainly distributed in Tiexi District, and its main risk source comes from industrial pollution.