My work examines how society influences and is impacted by machine learning and big data, with an increasing emphasis on data bias and data ethics. My current work examines content moderation, computer vision, and image recognition with a focus on racial, gender, and sexuality bias.
■ I am currently Co-PI on an NSF-funded study of case-based data ethics studies that will be ongoing or the next three years. We plan to examine how STEM students navigate ethical issues in key case studies and to identify better methods for inculcating data ethics into STEM students during their undergraduate studies before they enter the workforce.
■ My book project examines the heteronormativity of the computer vision-powered content moderation tools that currently detect and block ‘pornographic’ and ‘racy’ images online. I show how heteronormative bias creeps into our digital discourse, the worldview of coders, the code they write, the datasets they train that code on, and their decisions about how to adjudicate this bias after the fact.
■ Monea, A., Race and computer vision. In Sudmann, A. (ed.) The democratization of artificial intelligence: net politics in the era of learning algorithms 189–208. Bliefield, Germany: Transcript (distributed by Columbia University Press). (2019).