Background <p>Breast density is a significant risk factor for breast cancer and influences both the sensitivity and specificity of screening mammography. Mammographic interpretation can be challenging due to overlapping glandular tissue, leading to higher recall rates and false-positive findings. Advances in artificial intelligence (AI) have introduced tools that may assist radiologists in improving breast cancer detection, estimating breast density, and enhancing diagnostic performance by increasing sensitivity and specificity while reducing recall rates and interpretation time.</p> Objectives <p>To evaluate the performance of an artificial intelligence system in estimating breast density according to the American College of Radiology (ACR) classification using digital mammograms.</p> Methods <p>This retrospective study included 592 female patients who underwent full-field digital mammography (FFDM) in both craniocaudal (CC) and mediolateral oblique (MLO) views. Mammograms were independently assessed by two experienced breast imaging radiologists, blinded to each other’s results, for ACR breast density classification. All images were also analyzed using an AI-based software, and the results were compared with those of both radiologists.</p> Results <p>AI demonstrated an almost perfect agreement with researcher 1 (<i>κ</i> = 0.879) and a moderate agreement with researcher 2 (<i>κ</i> = 0.599) in classifying mammographic breast density.</p> Conclusion <p>Artificial intelligence demonstrated comparable performance to radiologists in assessing ACR breast density, showing strong potential for standardizing and automating density evaluation. Its integration into routine mammographic workflow may reduce inter-reader variability and improve reporting consistency.</p>