The ongoing artificial intelligence spring has brought many technological advancements, especially in generative artificial intelligence. Generative adversarial networks, diffusion models, and autoencoders can create synthetic yet highly realistic audio, photo, and video, which is referred to as deepfake. Building on the hypothesis that deepfake is a dual-purpose technology that can both support and impair human rights, this project investigates the positive applications of deepfake technology for prosocial purposes and two forms of its misuse: image-based sexual abuse and the weaponization of deepfakes to discredit journalism. Lastly, the project aims to develop a framework that maximizes the upsides of deepfake technology while minimizing its downsides.