Measuring bias in aggregated digitised content: A case study on Google arts and culture

Inna Kizhner, Melissa Terras, Maxim Rumyantsev, Valentina Khokhlova, Elizaveta Demeshkova

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Large cultural heritage aggregators, such as Europeana and Google Arts and Culture (GA&C), collect metadata and images from cultural institutions. They provide a single portal that introduces cultural heritage from around the world to the public (Sood, 2016, Petras et al., 2017). Selecting images and artifacts for these aggregators is an outcome of curatorial decisions, enlarging an art canon (Earhart, 2012, Feldman, 2016), building a cultural capital (Bertrand and Kamenica, 2018), and providing an infrastructure for a corpus of art history images (Drucker, 2013) that is critically important for the research in Digital Humanities. However, are such portals indeed a representative and balanced collection, the foundation for objective humanistic study and judgement? In this paper we argue that diversity, although present in GA&C, is too narrow to support our hope that it can act as a corpus of digital art history images. Our evidence proves that the digital corpus amplifies biases within the arts world towards western culture.
Original languageEnglish
Title of host publicationDigital Humanities 2019, Utrecht
PublisherAlliance of Digital Humanities Organisations
Publication statusPublished - 10 Jul 2019

Fingerprint

Dive into the research topics of 'Measuring bias in aggregated digitised content: A case study on Google arts and culture'. Together they form a unique fingerprint.

Cite this