Fellow 2024-2025

Bruno Moreschi's academic roots are in artistic research and museum collection digitisation, topics that have led him to his current research on Artificial Intelligence and Machine Learning. He explores critical and creative approaches to the use of new digital technologies, also transforming them into and through artistic, experimental, and collaborative methodologies. Moreschi's projects manifest themselves in artworks and exhibitions as well as in didactic formats and publications. In this way, he engages in institutional critique and brings the conditions and contexts of the use of Artificial Intelligence into the discourse of cultural studies. Moreschi is currently a Fellow at the Collegium Helveticum in Zurich, and at LIAS he will investigate the role of digital images in Large Scale Vision Datasets (LSVDs).

Abstract

Beyond First Glance: A Dive into AI/Computer Vision Through the Images in Large Scale Vision Datasets

Scientists often focus on the social impact of algorithms when studying AI/computer vision in society but fail to overcome the opacity that characterises their algorithms. Therefore, the project explores the background of machine learning and investigates machine vision using training data, the so-called Large Scale Vision Datasets (LSVDs). These datasets consist of millions of images taken without consent from social networks and organised through the precarious employment of thousands of microworkers. Analysing these images and the practices associated with them helps to understand that AI is not neutral but represents a historical continuity of dominant ideologies that existed long before the advent of computers. The scenes and processes to be organised correspond to the postcolonial logic of the distanced and extractive view, in an even more levelling way, due to the algorithmic scale. The study of these training datasets will draw on existing analyses, complemented by critical studies by other researchers focusing on AI inputs. The research will help to understand algorithms with greater accuracy and thus negotiate new, non-normative ways of perceiving the world through machines.

Education

2019 Ph.D., Arts, State University of Campinas, Sāo Paulo
2014 M.Sc., Arts, State University of Campinas, Sāo Paulo
2006 B.A, Communication, Federal University of Santa Catarina, Brasilien

Most Recent Academic Position

Early-career Fellow, Collegium Helveticum, the joint Institute for Advanced Studies (IAS) of ETH Zurich, the University of Zurich, and the Zurich University of the Arts

Most Recent Publications

„Five Experimentations in Computer Vision: Seeing (through) Images from Large Scale Vision Datasets.“ BJHS Themes 1 no. 17 (2023), 1–17.
with Gabriel Pereira. „Living with Images from Large-Scale Data Sets: A Critical Pedagogy for Scaling Down“, photographies 16 no. 2 (2023), 235–261.
mit Pereira, Gabriel, Andre Mintz, und Giselle Beiguelman. „We’ve Always Been Antagonistic: Algorithmic Resistances and Dissidences beyond the Global North.“ Media International Australia, Incorporating Culture & Policy 183, no. 1 (2022), 124–138.