PERIMETER

Digital assisted image, 2025 

Perimeter explores the mediation of personal history through machine learning. The series is entirely AI-generated and based on photographs of a factory where I once worked in my hometown—a site where photography was heavily restricted. Rather than reproducing existing images, the work attempts to reconstruct the photographs that were never taken: the unseen views, missed moments, and inaccessible spaces that remained outside the camera's field of vision.

Using photographic traces as a point of departure, the project treats machine learning as a tool for speculative reconstruction. The resulting images do not function as documentary records but as synthesized memories, generated from fragments of evidence and personal recollection. Suspended between fact and invention, Perimeter reflects on the gaps within photographic archives and the possibility of using AI to imagine what could not be photographed in the first place.