Shallow Fakes

Abstract

Data visualization technologies and production have bounded forward in recent years. At once the public’s appetite for data visualization in the news is increasing and barriers to making high quality visualizations — including in areas of cost, time, and education — are falling away. While exciting, these developments present concern. Specifically, gains in ease of use, accessibility, and enthusiasm over data visualization technologies increase the likelihoods of data visualization spoofing of bone fide news outlets and of manipulating or faking data. There has been much discussion of fake news due to its prevalence in the most recent U.S. presidential election and of the threat of “deep fakes” – images or video realistically doctored with machine learning technologies. There has been less discussion of data visualization fakes, which I’m calling “shallow fakes”. Yet, there is good reason to believe that a blight of “shallow fakes” could be on the near-horizon.