Penerapan Convolutional Neural Network untuk Membedakan Foto Sintetis Hasil Generative Artificial Intelligence dan Foto Asli

Abstract

The advancement of Generative Artificial Intelligence (GAI) has enabled the creation of synthetic images that closely resemble real photographs, raising challenges in verifying digital content authenticity. This study aims to develop a classification model based on Convolutional Neural Network (CNN) to distinguish between synthetic and real images accurately, as well as to implement it into a prototype detection system. The dataset consists of AI-generated and real images processed through preprocessing, training, and testing stages. Model performance is evaluated using accuracy, precision, recall, and F1-score. The results indicate that CNN effectively extracts visual features and achieves good classification performance. The model is then integrated into a web-based prototype to support real-time detection. The implementation results show that the system performs well; however, it still has limitations in handling unseen data outside the training distribution.