Leaking Subjects and Bounding Boxes
Author: Elisa Giardina Papa
Publisher: Sorry Press – Kubina Wiegand GbR
"Leaking Subjects and Bounding Boxes: On Training AI" examines how machines are trained to see, recognise and classify the world. Beginning in the winter of 2019, Elisa Giardina Papa worked as a "data cleaner" for several computer vision systems. Among the thousands of images she processed, she collected those that resisted the categories she was instructed to apply: ambiguous bodies, overlapping objects and forms that could not be neatly separated from their surroundings.
The book brings together these images and the instructions used to annotate them. Tasks such as outlining objects, identifying facial expressions and drawing bounding boxes reveal the human labour behind automated vision. They also expose the assumptions embedded in systems that divide images into discrete, recognisable and supposedly stable categories.
Rather than treating ambiguity as an error to be corrected, Giardina Papa considers what it might offer. Her collection turns attention towards bodies, desires and relationships that exceed classification, asking how we might see differently without making everything transparent, measurable or identifiable. The volume also includes a conversation between Giardina Papa and artist Mimi Onuoha, conducted by Zach Blas, on missing data and data cleaning.
Elisa Giardina Papa is an Italian artist and researcher born in 1979. She lives and works between New York and Sant'Ignazio, Sicily. Working across video installation, experimental film, writing and sculpture, she investigates forms of knowledge and desire that dominant systems of classification render illegible. Her practice addresses the relationships between gender, sexuality, care, labour and digital technologies, with particular attention to the human work underlying artificial intelligence.
She holds an MFA from the Rhode Island School of Design and a PhD in Film and Media from the University of California, Berkeley, where she completed her dissertation, The Incomputable Image: Untraining AI, in 2023. She is Assistant Professor of Modern Culture and Media at Brown University.
Her work has been presented at institutions including the Venice Biennale, the Museum of Modern Art and the Whitney Museum in New York, Gropius Bau in Berlin, the Secession in Vienna and M+ in Hong Kong. She is also a founding member of Radha May, an art collective formed with Nupur Mathur and Bathsheba Okwenje, whose work explores marginalised histories, gender, sexuality and colonial legacies.
- Year
- 2022
- ISBN
- 9783910265042
- Pages
- 376
- Price
- 28.04 EUR — In stock
From the same publisher
- Where Am I? Katalog (02)
- The Uncanny House
- Spirit Animal Animal Spirit
- Sorry Fax Only
- Simple Paintings
- Recalls Reconstructions Projections
Book categories
