The latest disinformation buzzword on everyone’s lips is “deepfake,” a term used to refer to videos that have been manipulated using computer imaging. (The word is a combination of “deep learning” and “fake.”) Using relatively inexpensive software, almost anyone can create a video whose subject appears to say or do something they never said or did. In one of the most recent examples, a Slovakian video artist named Ctrl Shift Face modified a video clip of comedian Bill Hader imitating Al Pacino and Arnold Schwarzenegger, so that Hader’s face morphs into that of the actors while he is imitating them. In another, a pair of artists created a deepfake of Facebook co-founder and CEO Mark Zuckerberg making sinister comments about his plans for the social network.
Technologists have been warning about the potential dangers of deepfakes for some time now; Nick Diakopolous, an assistant professor at Northwestern University, wrote a report called “Reporting in a Machine Reality” last year about the phenomenon. As the US inches closer to the 2020 election campaign, concerns have grown. The recent release of a doctored video of House Speaker Nancy Pelosi—slowed down to make her appear drunk—also fueled those concerns, although the Pelosi video was what some people have called a “cheapfake” or “shallowfake,” since it was obvious it had been manipulated. At a conference in Aspen this week, Mark Zuckerberg defended the fact the social network didn’t remove the Pelosi video, although he admitted it should not have taken so long to add
Read more here: https://www.cjr.org/analysis/legislation-deepfakes.php