This systematic review evaluates the diagnostic performance of Artificial Intelligence (AI) models designed for detecting caries lesions (CLs) in radiographic imaging. The review includes 20 articles published until January 2023, covering various imaging modalities such as periapical radiographs, bitewings, and orthopantomography. The studies used different AI algorithms, including artificial neural networks (ANNs), convolutional neural networks (CNNs), and deep convolutional neural networks (DCNNs). The diagnostic performance metrics, such as sensitivity, specificity, precision, and accuracy, were assessed, with results ranging from 0.44 to 0.98 for sensitivity, 0.85 to 0.98 for specificity, 0.50 to 0.94 for precision, and 0.73 to 0.98 for accuracy. Most studies exhibited a low risk of bias according to the QUADAS-2 evaluation. The review concludes that AI-based models show good diagnostic performance in CL detection, potentially aiding dentists in early and accurate diagnosis, reducing healthcare costs, and improving patient care. However, limitations include the size and heterogeneity of datasets, and future research should focus on larger, more standardized datasets to enhance the reliability and generalizability of AI models in dental caries detection.This systematic review evaluates the diagnostic performance of Artificial Intelligence (AI) models designed for detecting caries lesions (CLs) in radiographic imaging. The review includes 20 articles published until January 2023, covering various imaging modalities such as periapical radiographs, bitewings, and orthopantomography. The studies used different AI algorithms, including artificial neural networks (ANNs), convolutional neural networks (CNNs), and deep convolutional neural networks (DCNNs). The diagnostic performance metrics, such as sensitivity, specificity, precision, and accuracy, were assessed, with results ranging from 0.44 to 0.98 for sensitivity, 0.85 to 0.98 for specificity, 0.50 to 0.94 for precision, and 0.73 to 0.98 for accuracy. Most studies exhibited a low risk of bias according to the QUADAS-2 evaluation. The review concludes that AI-based models show good diagnostic performance in CL detection, potentially aiding dentists in early and accurate diagnosis, reducing healthcare costs, and improving patient care. However, limitations include the size and heterogeneity of datasets, and future research should focus on larger, more standardized datasets to enhance the reliability and generalizability of AI models in dental caries detection.