Deepfake images created using artificial intelligence have become a serious challenge in the modern digital world. These images are generated or manipulated using advanced AI technologies to produce highly realistic fake visual content. The misuse of deepfake images may lead to misinformation, identity theft, cybercrime, social manipulation, and digital fraud. As AI-generated images become more realistic, it becomes difficult to distinguish fake images from genuine images through normal visual observation. Therefore, digital forensic analysis plays an important role in identifying manipulated or AI-generated images. This study focuses on the AI-based detection and forensic analysis of deepfake images using different forensic examination techniques. A total of twenty sample images, including ten real images and ten AI-generated images, were analyzed using AI image detection tools, Error Level Analysis (ELA), clone detection, and metadata analysis. The forensic examination was conducted to identify visual inconsistencies, editing traces, image manipulation artifacts, and metadata variations between real and AI-generated images.