We Have Always Been Biased

This paper aims to provide a historical context in which to appreciate the shared problems faced by social and computer scientists in using and creating biometric data. While historians of science are well aware of the trials (and errors) of previous attempts to quantify the human condition, this literature has not always made it into discussions of modern biometrics. Indeed, manuals for what are now called the computational social sciences often imagine that data mining and statistical averages are new, and that “Big Data” has only existed for the past decade. Such historical amnesia has led, this paper argues, to problems of modern bias emerging in the literature as a technical issue rather than a full-fledged conceptual barrier with long roots. Seen only in the light of present politics and practical concerns, I argue that these problems will remain intractable.