“The Machine That Remembers Without Knowing”: Copyright, Artificial Cognition, and the Crisis of Authorship under Zambia’s Copyright and Performance Rights Ac

The rapid emergence of large language models (LLMs) has destabilised foundational assumptionsin copyright law, particularly concerning authorship, originality, and infringement. This articleinterrogates the adequacy of Zambia’s Copyright and Performance Rights Act, Chapter 406 of theLaws of Zambia (Act No. 44 of 1994, as amended in 2010), in addressing the challenges posed byartificial intelligence systems trained on copyrighted materials. Moving beyond the conventional fairuse discourse, the article advances a doctrinal critique centred on the concept of “non-humanreproduction without cognition,” arguing that existing legal frameworks are ill-equipped to regulatemachine-mediated copying that neither fits traditional infringement nor recognised exceptions.Drawing on Zambian statutory law, common law principles, comparative jurisprudence, andcontemporary policy developments, the article proposes a reconceptualisation of copyrightenforcement grounded in functional harm and technological neutrality.