This article critically assesses the adequacy of South African performers' rights in addressing the legal challenges posed by artificial intelligence (AI), with a particular focus on deepfakes – synthetic media created through machine learning techniques, falsely depicting people doing or saying things that they did not do or say. While the Performers' Protection Act 11 of 1967 has been proposed as a potential tool for addressing some of the harms associated with deepfakes, this legal framework was not designed to confront the technological and normative challenges introduced by such AI-generated content. The article argues that performers' rights are conceptually ill-suited to regulate synthetic digital imitations, especially where no protected performance has taken place. Consequently, the current legal regime offers insufficient recourse to performers whose identity is simulated or distorted by AI-generated content. Moreover, individuals who do not fall within the statutory definition of "performer", such as ordinary citizens, are left with virtually no recourse under this regime, despite being equally vulnerable to reputation and personal harm. By exposing these doctrinal limitations, the article highlights the broader inadequacy of existing South African law in responding to the challenges posed by synthetic media. It concludes that a reconfiguration of the legal approach is required, one that more directly engages with the realities of AI-generated expression and its legal implications.
G. M. Jansen (Tue,) studied this question.