beermoney, 9 channels (36s each), video, no sound, 2019

Semantic segmentation is an artificial intelligence process of identifying different areas in a still image and attaching to them a class label. In order to train and refine this process of discrimination, areas in these photos, are manually identified, outlined and labeled by free-lance workers. In online forums, these kinds of tasks are commonly referred to as beermoney jobs––a source of income devoted exclusively to leisure.  This artificial intelligence technology was initially created for military application and is currently in development for use in the film industry to quickly identify elements in a moving image and replace, change or erase them.

For this project, I individually uploaded the 878 frames of the 36 seconds film Employees Leaving the Lumière Factory, 1895 version, onto an online platform, which assigned those images randomly to online freelance workers around the globe. Each frame was identified, outlined and labeled individually by different workers. The class label system was defined upon the guidelines suggested by the website, in this case: person, pavement, sidewalk, wall, door, background, animal, vehicle and glitch.

After six months all the 878 frames were received back. By using the label system they were re-assemble into 9 videos of 36 seconds each–each video represents what the machine would see in order to identify, erase or replace any of those categories.