Algorithm Datalogic : Audits, Accountability and Diverse Training Data through Interdisciplinary Social Science

Abstract

The datalogic of contemporary urban societies and overreliance on data mining through sensors and platforms supported by digital infrastructures could be considered foreboding. Data-driven form of governance constitutes the most current turn into a legitimized mode of knowledge acquisition and epistemic development at all its complex levels. Algorithms play a crucial role at the intersection of data and analysis. This paper critically examines the role of algorithms including their technical and methodological capacities, that serve to manage data across many different services, products, and public and private organizations. Furthermore, the use of data across disciplinary boundaries generates a form of knowledge acquisition that is increasingly reductive even though, paradoxically, is based on expansive and vast amounts of data. Arguing for diverse cognitions, epistemologies, and conceptualizations of data, this study examines algorithm bias and audits to maintain accountability for the training data mainly used by AI. The paper concludes with plausible recommendations for future interdisciplinary social science research within the convergence science paradigm to demonstrate a critical understanding of the current stage of development of artificial intelligence and the social.

Presenters

Julia Nevarez
Professor, Sociology, Kean University, New York, United States

Details

Presentation Type

Paper Presentation in a Themed Session

Theme

Social and Community Studies

KEYWORDS

Sociology, Technology, Inequality