Data-driven hiring systems may incorporate unfairness

In the 2014 report "Networked Employment Discrimination", the Future of Work Project studied data-driven hiring systems, which often rely on data prospective employees have no idea may be used, such as the results of Google searches, and other stray personal data scattered online. In addition, digital recruiting systems that only accept online input exclude those who do not have internet access at home and must rely on libraries and other places with limited access and hours to fill in and submit applications. A 2009 survey found that 16% of retail hiring decisions also used pre-employment psychological, personality, cognitive, and other tests even though these may not reflect how prospective employees will perform in the job they're applying for. The report concludes that while employers have access to far more comprehensive profiles of candidates than ever before, it is unclear what they're optimising for or whether the process they're using is fair. The lack of transparency makes it possible to mask hiring processes that contain prejudicial elements.

https://www.datasociety.net/pubs/fow/EmploymentDiscrimination.pdf

Writer: Alex Rosenblat, Tamara Kneese, and Danah Boyd

Publication: Data & Society Research Institute