Cart Search

2024 Annual Conference

St. Louis

March 21-23, 2024

submit Remember my login

Looking for a chapter event?

Past SPE Annual Conferences

Kaitlyn Jo Smith

SPE Member since 2016
Member Chapter: Southwest

Lights Out, 2020

In Lights Out, I utilize machine learning to render visible the intangible state of America's labor force. The term lights out manufacturing refers to a fully automated factory that requires no human presence on site, operating without heating, air conditioning, lunch breaks or unions. Similarly, the portraits in Lights Out are not real people, rather, they are deep fakes created by artificial intelligence utilizing a neural network. This neural network was trained using a dataset of 50,000 pictures of factory workers that I sourced through facebook. I empower the machine to visualize those it has replaced by employing these found identities to teach the algorithm to see, interpret and render out its own assembly of laborers. As the neural network is fed more portraits to analyze, its deep fakes become more believable, resulting in a new labor force of 60,000 - one worker for each factory shuttered since 2000. The AI generated faces have been meticulously organized into columns that slowly shift up and down the gallery wall. The monotonous, machine-like rhythm created by this vertical assembly line mimics that of a conveyor belt, symbolically placing the portraits in dialog with capitalist commodities.

Lights Out

Lights Out

Email Sign Up

SPE email updates contain resources, news, and more!

About this piece

Comments about this piece

Dialogue and critique are important to the SPE mission.
Please join the conversation.

Exit Full Screen Mode