M.A.I

M.A.I M.A.I. (Merging Archival Interpretation) explores archival processes by combining human interpretation with AI-driven expansion. In collaboration with Archiv der Moderne, the project uses photographs of archived collectibles as visual references to generate new graphic elements. This creates a dialogue between the original objects and their speculative reinterpretations. The experience was presented as an interactive installation at the Summaery 2023 (07/13/23 – 07/16/23) in Weimar. Visitors were invited to engage directly with the archive's potential for endless transformation. The process began with the selection of a card, each representing a piece of the archive, a fragment of memory. By scanning the chosen card, a dialogue with the artificial intelligence, specifically OpenAI's DALL-E 2 model, was initiated. From there, one could simply watch the endless generative expansion unfold as the AI began to generate new images based on the selection. This process could theoretically continue indefinitely, a perpetual reinterpretation of the past. To interrupt the cycle and begin a new visual exploration, a new card had to be selected. The resulting videos are documents of these individual interactions, their lengths determined by how long each participant allowed a single card to generate before intervening. They are a record of human choice and machine speculation. With the use of AI, these elements are further enriched, blending traditional archival practice with generative technologies. In doing so, M.A.I. also reflects on the inherent loss within archival work: gaps, omissions, and forgotten narratives. AI becomes both a tool of reconstruction and imagination, making absence visible. M.A.I. offers a new perspective on how archives may be processed in the future, merging past and future, presence and disappearance, human memory and machine speculation.