Models of seeing

Photographic series, 2025. Print on glass.

Models of Seeing explores the friction between human vision and AI computer vision models. A collection of historical photographs has been analyzed using YOLO9000, a computer vision model that has been trained using a dataset of 9,000 words and millions of images. What does YOLO9000 see? AI models have a compulsive side; for them, seeing means to project meaning without rest. Labels are automatically assigned and a predetermined worldview is overlapped onto images. This collection of photographs shows glimpses of when this operation fails due to a poetic interference. The goal of data extraction is not achieved and another model of seeing emerges, one in which the meaning of images is not closed but opened up by words.

The series was exhibited at the OFFSCREEN fair in Paris (2025), held at the Chapelle Saint-Louis de la Salpêtrière, with Chiquita Room gallery.