Train Connect Assemble Extract, Video, 2:50 min
Departing from Antje Ehmann’s and Harun Farocki’s "Labour in a Single Shot" the video highlights the invisible contribution of workers around the globe that build the computational infrastructure. Inspired by a lecture of Kate Crawford at MIT in February 2023, and her book "The Atlas of AI", the video zooms out and tries to visualize the human realities that underpin the work at MIT. The video addresses four labor aspects of artificial intelligence often overlooked: Training: Labeling and content moderating behind computational intelligence, Connecting: establishing a global infrastructure from Wi-Fi routers to transatlantic network cables, Assembling: producing components for contractors of US tech companies, and, Extracting: Mining earths, such as quartz sand for the creation of silicon dioxide, which is needed for computational components such as CPUs.
For lack of documentary visual material, I asked Dall-E, the deep learning model of OpenAI, how it would imagine this labor that is so fundamental to its own existence. Prompts for the AI included "A geometric view of an infinite number of content moderators in a dark office staring at blue screens", "central view, hundreds of divers lined up holding network cables, underwater, green light", "A central view of hundreds of workers in a factory in Indonesia assembling laptops", and "The grim reality of African workers who mine rare earths". The video, in a way, is a conceptual mise en abyme. The images created by Dall-E let us see the invisible reality of human labor without which neither the artificial intelligence nor the underlying work would exist.
Time
2023
Client
Self-initiated / Massachusetts Institute of Technology
Place
Cambridge, USA
Type
artwork