No Blacks: How algorithmic “white goodness” pushes Black job seekers out of online recruitment pools

• In Event: 20220 - Section on Communication, Information Technology, and Media Sociology Roundtables
  In Refereed Roundtable (60 minute): Section on Communication, Information Technology, and Media Sociology Roundtable X: Social Media II

Abstract

Social media platforms offer employers the ability to present their vacancies to their preferred audience. Through a process called targeted advertising.(Nadler et al, 2018; Basileal et al, 2021; Zuboff, 2019). Thereby relegating the work of interpreting “first impressions” to a network of algorithmic systems, which decide which candidate may be the best fit for a job (Nadler, 2018; Zuboff, 2019). This objective measure of employer/employee compatibility may seem like a benefit to Black job seekers. However, this is not necessarily the case. There is a growing body of research that finds offline racial discrimination migrates to the digital world. This is a phenomenon called algorithmic bias (Buolamwini and Gebru, 2018; Sweeney, 2013). A matter examined by a team of computer scientists in the 2021 paper “Annotators With Attitude”, which found automated content moderation systems were more likely to tag words and phrase used in African American Vernacular English (AAVE) as toxic than messages written in standard English (Sap et al, 2021). This negatively impacts Black job seekers with “Black: sounding names, because they are more likely to be tagged negatively, because they are not written in standard English. And these candidates are therefore more likely to be pushed out of the recruitment pool (Wilson et al, 2024). This paper contends these examples of algorithmic bias invokes W.E.B Dubois’s theory of white preference, made in his 1910 book “Dark Water”. In which he states American society “assumes that all hues of God’s whiteness alone [are] inherently and obviously better than brownness or tan,” (Dubois, 1910). This study seeks to test this hypothesis by interrogating the epistemological drivers of the platform policy formation process. By analyzing the findings of a series of in depth interviews with trust and safety leaders across a number of Silicon Valley based social media companies.

Author

• Mutale Nkonde, University of Cambridge