Trevor Paglen

I think about clouds a lot. They’re really strange things – on one hand, humans have gone to great lengths to characterize different kinds of clouds, creating intricate cloud atlases and taxonomies. On the other hand, when we look at clouds we tend to...

When I was working on the Landing Sites images, it occurred to me that if one were to literally “dive into” the seascapes that I was making images of, one should theoretically be able to see the actual material conjunctions of internet cables that I...

This is an ongoing project to track and photograph the world of secret satellites that are above our heads nearly constantly.
Although the satellites I track and photograph in this series are not typically acknowledged by the United States, they are cataloged and observed by an international network of amateur “satellite observers,” whose observations I use to calculate the...

ImageNet Roulette was part of a broader project to draw attention to the things that can – and regularly do – go wrong when artificial intelligence models are trained on problematic training data.
ImageNet Roulette is trained on the “person” categories from a dataset called ImageNet (developed at Princeton and Stanford Universities in 2009), one of the most widely used training sets in machine learning research and development....