Plumbing the depths of deepfakes: What’s at stake, what’s to be done, was prepared with the support of Research ICT Africa and CETIC (Brazil), as Knowledge Partners to the G20 Digital Economy Working Group (DEWG). Developed under Priority 4: Equitable, Inclusive and Just AI, this Issue Brief offers contemporary considerations on the sustainability of fact-based information environments and online information integrity in light of generative AI (GenAI).
This brief addresses the urgent challenge of developing appropriate policy responses to GenAI in a fast-moving technical context. It underscores the value of foresight and research in detecting the instances and extent of deepfakes and highlights significant empirical evidence demonstrating GenAI’s impact on social trust, human rights, and economic relations. Building on the DEWG’s work on information integrity undertaken during the Brazilian G20 Presidency in 2024, the Issue Brief provides recommendations, including using scenario planning and human rights risk assessments to mitigate GenAI risks and developing transparency and disclosure protocols for state actors.
RIA’s contribution to this work reflects its substantial policy research and engagement under the Information Integrity and Africa Just AI projects, both of which address AI’s impacts on our threatened information ecosystem and the many disadvantages it poses to vulnerable groups, including women, children and marginalised voices. The resultant recommendations are cross-cutting and focus on data governance as an important element of GenAI risk mitigation. It calls for coordinated efforts and task forces, as well as harmonised national technical standards for content provenance, among other measures.
