In 2018, legacy car manufacturers such as Mercedes, BMW, and Audi, taken aback by the advents of Google and Tesla, invested a lot of money in research on autonomous vehicles. They gathered an immense amount of images taken by their cars that they needed to process in order to train their algorithms. This process, called "image segmentation", consists in manually outlining and labelling elements of interest in the image. It is very labor intensive and still cannot be automatized, it is thus outsourced to online micro-workers from the Global South.
Unknown Label explores the daily reality of online micro-workers from Venezuela, Kenya, and the Philippines who annotate images for self-driving cars. It investigates the power asymmetries and neocolonialist exploitation involved in the human labor necessary to train AI systems, but also the many micro-gestures of resistance that workers share on private groups.
Unknown Label reveals the hidden people that help shape how machines see the world, as well as the categories used to encapsulate our world and make it accessible to the machines, thus asking questions about the politics of these categories and to what extent they in turn affect the world we live in.
In 2018, legacy car manufacturers such as Mercedes, BMW, and Audi, taken aback by the advents of Google and Tesla, invested a lot of money in research on autonomous vehicles. They gathered an immense amount of images taken by their cars that they needed to process in order to train their algorithms. This process, called "image segmentation", consists in manually outlining and labelling elements of interest in the image. It is very labor intensive and still cannot be automatized, it is thus outsourced to online micro-workers from the Global South.
Unknown Label explores the daily reality of online micro-workers from Venezuela, Kenya, and the Philippines who annotate images for self-driving cars. It investigates the power asymmetries and neocolonialist exploitation involved in the human labor necessary to train AI systems, but also the many micro-gestures of resistance that workers share on private groups.
Unknown Label reveals the hidden people that help shape how machines see the world, as well as the categories used to encapsulate our world and make it accessible to the machines, thus asking questions about the politics of these categories and to what extent they in turn affect the world we live in.
Testimonies: Oliver, Elvina, Ivon, Yonaille, Jonel
Research assistants: Leonard Nally Simala, Andrea Paola Hernandez, Niside Panebianco
Editing: Lucas Azémar & Nicolas Gourault
Screenwriting advisor: Quentin Faucheux
Image assistant: Héléna Michaud
Sound editing & re-recording: Etienne André
Thanks to: Inès Sieulle, Antoine Chapon, Claire Williams, Clara Chapus, Charlotte Cherici, Yuyan Wang, Peter Zorn, Marcie Jost, SoundImageCulture, Amir Borenstein, Mary Jimenez, Jorge Leon, Effi Weiss, Julie Pfleiderer, Jan De Coster, Susanne Weck, Julien Chapelle, Sasha Litvintseva & Beny Wagner, Don Quichotte Films, Yannick Beauquis, Quentin Brayer, Julian Posada, Milagros Miceli, Paola Tubaro, Sana Ahmad, Alex Roy, Phil Koopman, Adrien Gaidon, Winifred Poster
This work was realized within the framework of the European Media Art Platform residency program at Werkleitz with the support of The Creative Europe Culture Programme of the European Union.
This film was developed in the framework of SoundImageCulture with the support of the Wallonia-Brussels Federation, Vlaams Audiovisueel Fonds and Vlaamse Gemeenschapscommissie
With support from: EMAP European Media Art Platform Residency Programme 2023; SoundImageCulture 2023
Nicolas Gourault (FR) is an artist and filmmaker based in Paris with a background in visual arts and visual studies. He has worked with Forensic Architecture before graduating from Le Fresnoy – Studio national des arts contemporains. His work is imbued with this double training, navigating between online investigations and the use of new media as documentary tools. He explores the power relationships embedded in technologies and builds counter-narratives through testimonies and experimental image-making. His artworks have been exhibited in contemporary art venues but also in film festivals.
Nicolas Gourault (FR) is an artist and filmmaker based in Paris with a background in visual arts and visual studies. He has worked with Forensic Architecture before graduating from Le Fresnoy – Studio national des arts contemporains. His work is imbued with this double training, navigating between online investigations and the use of new media as documentary tools. He explores the power relationships embedded in technologies and builds counter-narratives through testimonies and experimental image-making. His artworks have been exhibited in contemporary art venues but also in film festivals.
Unknown Label captivated the jury with its innovative approach to documentary storytelling, seamlessly integrating animation as the core narrative medium and redefining the documentary film format. The animation is more than mere illustration, it is integrated as the very essence of its narrative; the documentary could not exist without it, and it is an exemplar for the New Animation Art category of Prix Ars Electronica. The film opens with a simple, single-color animation illustrating the process of segmentation, introducing viewers to the unseen individuals who label training data for AI vision systems. As the story delves deeper into the complexity of their work, social dynamics, and the discrimination faced by these workers, the animated visuals evolve in complexity as well. The narrative crescendos with a city-scale 3D data visualization, revealing the staggering amount of invisible human labor that goes into training our AI systems. The documentary navigates complex themes with clarity and empathy, elevating the discourse on AI ethics and globalization.
Unknown Label captivated the jury with its innovative approach to documentary storytelling, seamlessly integrating animation as the core narrative medium and redefining the documentary film format. The animation is more than mere illustration, it is integrated as the very essence of its narrative; the documentary could not exist without it, and it is an exemplar for the New Animation Art category of Prix Ars Electronica. The film opens with a simple, single-color animation illustrating the process of segmentation, introducing viewers to the unseen individuals who label training data for AI vision systems. As the story delves deeper into the complexity of their work, social dynamics, and the discrimination faced by these workers, the animated visuals evolve in complexity as well. The narrative crescendos with a city-scale 3D data visualization, revealing the staggering amount of invisible human labor that goes into training our AI systems. The documentary navigates complex themes with clarity and empathy, elevating the discourse on AI ethics and globalization.