97 Things About Ethics Everyone in Data Science Should Know

Chapter 49. Avoid the Wrong Part of the Creepiness Scale

Hugh Watson

Professor of MIS, Terry College of Business, University of Georgia

Some algorithms that make use of personal data are perceived as helpful, such as movie recommendations on Netflix, recommendations for nearby restaurants from Yelp, and Waze’s driving routings. Others are creepy. Remember the first time you viewed an item on the internet and ads from various vendors followed for days? It’s also creepy to meet someone for the first time and soon thereafter receive a Facebook friend recommendation for that person. Some uses of personal data are just so wrong, such as Russia’s attempts to influence election results through targeted news feeds, or you receiving an ad from a reseller of engagement rings after changing your relationship status from “Engaged” to “Single” on Facebook.

The different reactions to the use of personal data and algorithms can be considered using a creepiness scale, as shown in the following figure, where the degree of creepiness is on the y-axis and the extent of use of personal data and algorithms is on the x-axis.

Something is creepy when it differs from the norm and is perceived to be potentially threatening or harmful. To illustrate, an app that reveals your location to others without ...