<p>The proliferation of deep learning technologies has led to the emergence of highly realistic deepfake face-swap content, where one person’s identity is convincingly replaced with another in videos or images. This manipulation poses a serious threat to personal privacy, biometric security, and public trust, making it a powerful tool for misinformation, identity theft, and social engineering. As deepfakes continue to evolve in quality and accessibility, developing effective detection and prevention methods has become a critical research priority. This survey presents a comprehensive overview of the current landscape of deepfake face-swap generation and detection techniques. We analyze recent advancements in generative models, including Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and Diffusion Models. On the detection front, we categorize and evaluate image- and video-based approaches leveraging convolutional neural networks, frequency-domain cues, and multi-modal fusion strategies. The paper also examines the limitations of existing methods in addressing real-world challenges such as compression artifacts, low resolution, and adversarial attacks. Benchmark datasets and performance metrics are critically reviewed. In conclusion, we highlight promising research directions–including blockchain-based verification and hybrid detection frameworks–to strengthen the integrity of digital media and provide a roadmap for future advancements in this evolving field.</p>