Treating data as a living cultural archive, these visual investigations confront representational gaps and epistemic harms in AI with regards to the SWANA region. The project responds to contemporary datalogical, technical, and aesthetic biases that misrepresent or Arab identities, and creates new alternatives for self-representation that are at simultaneously archival, nostalgic, and playful. These generative archives, built using LoRA models trained on mid-cenury records, create possibilities for past and future to intermingle as digital artifacts.
