<![CDATA[The rapid advancement of artificial intelligence (AI) technology has led to an increase in AI-generated media that is increasingly difficult to distinguish from authentic media. This phenomenon presents significant challenges in various fields, including digital security and the creative industry, necessitating a reliable automatic detection system. Manual identification is inefficient and error-prone, as it requires specialized expertise to recognize subtle digital media. This study proposes the development of an AI-generated media detection system using a Convolutional Neural Network (CNN) with the ResNet-50 architecture. The ResNet-50 model was chosen due to its proven ability to handle deep feature extraction and overcome the vanishing gradient problem. A sample of 4.600 images, 2.300 AI images and 2.300 real images, were used. The research methodology included data collection from various Kaggle datasets, data preprocessing including resizing and augmentation, model training, and performance evaluation based on accuracy, precision, and recall metrics. Experimental results show that the developed model achieved a high testing accuracy of 97.30%. This indicates the model's capability to effectively classify media as AI-generated or real with high precision. This research is expected to be a valuable reference for the development of more accurate and efficient AI systems in detecting synthetic media.]]>
Copyrights © 2026
