Bovina, NY


2018-2019

This work trains a neural network on a single location—County Highway 5 in Bovina, NY—to explore how machines construct understanding of landscape from limited observation. Rather than the millions of images typically used, the system learns from a constrained set of rural photographs, developing its own visual language for representing natural spaces.

The resulting images are not accurate, but accuracy isn’t the objective. The network creates internally consistent landscapes that follow its own organizational logic rather than photographic fidelity—synthetic environments that exist as computational interpretations rather than documentary records. What emerges reveals how artificial intelligence builds meaning from visual data: not through direct perception, but through pattern and inference.

The work invites encounter with an alien yet strangely familiar way of seeing—one that echoes the history of artists who have worked in this same Hudson Valley landscape while operating according to principles entirely its own.

For more on the concepts behind this work, see the essay Little AI.

Learning Nature • Winter Woods • Cloud Canyon • Dandelions - AI / machine learning generated images and videos

Little AI - AI / Machine learning sculpture