Facebook banned the highly altered, algorithmically generated videos known as deepfakes from its platform last week. But what may be a bigger problem for our politics has a less menacing name: cheapfakes.
There's no question deepfakes are frightening. It's possible for anyone, with some money to spare, to use machine learning to simulate an individual saying something they never said or doing something they never did--and difficult for those rooting out the dissimulation to stay ahead. The Washington Post reports that start-ups are selling images of false faces to companies that want to increase diversity in their ads and even to a dating app that wants the pictures for a chatbot.
Yet deepfakes have hardly flooded our political conversation with synthetic President Donald Trumps singing the praises of the Green New Deal or synthetic Elizabeth Warrens giving three cheers for corporate corruption. They're far more common in the realm of pornography, which is already prohibited on popular platforms. What have flooded our conversations are low-effort edited versions of existing video: that viral clip from last year of House Speaker Nancy Pelosi, D-Calif., distorted so she appeared drunk, or a similarly popular contextless snippet from just last week of Joe Biden seeming to parrot a white-supremacist talking point.
These are also examples of media fiddled with deliberately to fool people and leveraged as digital smears against the individuals they depict. Yet because Facebook's policy affects only video that is the product of artificial intelligence or machine learning, they don't violate its latest set of rules. Instead, these cheapfakes get similar treatment to traditional misinformation such as made-up news stories: demoted in users' feeds if a third-party fact-checker rates them as false and, in an important update, labeled with a warning screen. But they are not banned.
It makes sense that Facebook would rather forbid a single, scary technology than take on the more complicated category of manipulated media generally. Drawing lines, and distinguishing between legitimate parody and harmful lie, will be difficult. But to confront the specter that actually faces our democracy today rather than just a ghost of deception to come, all platforms need to accept the larger, more difficult challenge. That will mean examining the entire spectrum from gently to egregiously edited--and focusing on a video's effect rather than the method of its production. Otherwise, our democracy will soon discover that cheapfakes come at great cost.
Editorial on 01/11/2020
Print Headline: Cheapfakes