Democratising AI through Culture

This report explores the possibility to make AI more inclusive and participatory through the concept of “AI as a cultural commons”. It proposes a practical approach for cultural practitioners to decolonise current AI systems by influencing representation in generative AI. Through a case study of South Korea, it presents a perspective on AI development outside North America and Europe, and demonstrates how AI development is heavily influenced by (and reinforces) patriarchal norms, operating with neo-colonial logic. The authors propose how lay users/non-tech experts can be empowered to critically reflect on the digital space through a feminist and decolonial lens, and developed a toolkit to enable non-experts to experiment and locally train AI models to generate alternative more inclusive narratives, compared to the version reproduced through biases inherent in existing commercially available AI models. What is “AI as a cultural commons”? What policies and social infrastructure need to exist to make AI development more inclusive and participatory through this approach? How can actors and stakeholders in international cultural relations (ICR) do to make “AI as a cultural commons” a reality? What concrete initiatives and ideas can policymakers, cultural institutions, and activists for digital rights implement to support participatory processes in AI design and development?