Abstract

The right to identity aims to protect the subjective interests of individuals in their likeness, image, voice, and other distinctive personality attributes. The right to identity is legally recognised in South Africa, but deepfakes have a tendency to devalue this right. Deepfakes are created with deep learning software that enables users to create deceptive videos, sound recordings, and photographs of events and people that are indistinct from reality. This goes against a person’s right to control the use of their likeness. South African law does not directly regulate the creation and publication of deepfakes. Liability for the publication of deepfakes may be established using principles in different fields of law, such as the law of delict and criminal law. However, the dissemination of deepfakes on the internet continues to evolve, as they become more difficult to detect, and this necessitates a new perspective on how to provide sufficient remedies for victims whose right to identity is violated through deepfakes. It also calls for the refinement of establishing the liability of people who are tagged to deepfakes posted on social media. This article aims to highlight the challenges in protecting the right to identity and establishing liability under South African law in the context of deepfakes.