When Is an Algorithm Transparent?: Predictive Analytics, Privacy, and Public Policy
IEEE: Security & Privacy, May/June 2018
9 Pages Posted: 12 Oct 2017 Last revised: 21 Sep 2018
Date Written: October 12, 2017
The rise of data mining and predictive analytics makes the problem of algorithmic transparency pressing. Solving that problem requires answers to two questions. What are the criteria of transparency? And how do you tell whether a predictive system meets those criteria? We confine our attention to consumers engaged in commercial transactions because this already raises most of the questions that concern us. We propose that predictive systems are transparent for consumers if they able to readily ascertain the risks and benefits associated with the predictive systems to which they are subject. We examine three ways to meet this condition: disclosing source code; techniques that reveal how an algorithm works without disclosing source code; and reliance on informational norms.
Keywords: transparency, algorithms, transparency of algorithms, privacy, public policy, predictive analytics, norms, informational norms
JEL Classification: C80, K29
Suggested Citation: Suggested Citation