Governing artificial intelligence under the African human rights system : drawing lessons from international best practices

Abstract

This study explores the state of governance of Artificial Intelligence under the African human rights system. The study finds that AI is currently regulated by the African Charter and a variation of soft law instruments such as the Declaration of principles on free expression and access to information and Resolution 473 of the African Commission. However these protections are minimal as most of them predate the AI age and are not tailor made to address the contemporary challenges of a fast-automating world. The study finds merit in the AI regulatory models under the European Union and the Council of Europe and recommend that the African Commission, STC on ICTs and AUCIL garner Best practices from these global standard setting bodies. Among these best practices are vigorous multi-disciplinary preparatory research, a risk based and precautionary approach and generating human rights based legal standards at every stage of the algorithmic lifecycle.