Working the Image

Organized by: Bruno Moreschi and Gabriel Pereira
Venue: Collegium Helveticum

⇒ This is a public event. Participation is free of charge and registration is not required.
⇒ Virtual participation via this link

In this workshop, Gabriel Pereira from the University of Amsterdam and the artist and Early-Career Fellow Bruno Moreschi will coordinate exercises that critically discuss the association between visual content and metadata, tagging processes and collaborative ecosystems to expand the informational power of images from machine learning datasets.
Sonia Dantas and Julio Kraemer will participate virtually to share their experience as crowdworkers. They both perform discrete on-demand tasks on platforms such as Amazon Mechanical Turk that computers are currently unable to do as economically. This includes task such as identifying specific content and tagging images accordingly. Sonia and Julio are two of the five crowdworkers who are part of the free print publication Exch w/ Turkers that will be launched. They and the publication’s designer Guilherme Falcão will participate in a panel about the publication that summarizes exchanges with people who train artificial intelligence models.

Image credit: The decanonization process in one image of the Open Image dataset. The white rectangle corresponds to the area of the image that Google’s AI considers relevant. Credits: Bernardo Fontes and Bruno Moreschi, GAIA–C4AI, InovaUSP.