Abstract
Artificial intelligence (AI) systems are increasingly becoming a part of everyday life, and with that comes a rise in the generation of creative outputs with minimal human intervention. Therefore, the boundaries that exist for authorship and originality within the scope of copyright law are challenged. This minor dissertation critically analyses and examines what the legal status of these AI-generated works is within the context of South African copyright law, particularly in reference to the Copyright Act 98 of 1978. It interrogates whether these works meet the statutory requirements of ‘authorship’ and ‘originality’ and whether the existing legislation can accommodate AI-generated works, as South African legislation is human-centric, as there is currently no provision for AI-generated works in the Copyright Act.
The analysis is structured into three key dimensions, namely: the interpretation of ‘authorship’ under South Africa’s Copyright Act and its implication for AI-generated works; the threshold for ‘originality’ in works that are produced by AI or in cases where there is assistance of generative algorithms; the policy and doctrinal consequences of extending or withholding copyright protection from these works. Foreign legislation (US and UK) is used to draw comparisons, together with international instruments (TRIPS Agreement and Berne Convention). These are used to highlight the alternative approaches to ‘authorship’ and ‘originality’ by offering broader context for potential reform.
Ultimately, this minor dissertation argues that South African copyright legislation fails to adequately address the emerging technological realities of the rise of AI-generated content by placing emphasis on human intellectual effort. Therefore, re-evaluation of the definition of ‘authorship’ is required, as well as a principled and forward-looking approach to reform, in order to ensure legal certainty while promoting innovation and aligning copyright doctrine with global technological developments.