GAN Carpet Project / Forbidden
Machine-learning pattern works drawn from Persian, Azeri and Armenian carpet traditions.
The GAN Carpet Project grows out of my long interest in the relationship between digital images and woven form. Working with carpet and tapestry traditions through machine learning systems creates a complicated field of pattern, cultural memory, authorship and translation.
The project is also part of NPT’s larger concern with Iran, Azerbaijan, Armenia, the UAE and the mediated histories that surround regional pattern and image-making. When AI recombines carpet traditions, it produces beauty, but it also raises questions about appropriation, memory, dataset, loss and abstraction.
For me, these works connect the Jacquard tapestries to AI. The woven object is not the end point of the digital image; it is a place where computation, textile history and cultural pattern become visible at once.