Madeleine Kearns of National Review Online describes the growing threat of a problem rooted in technological prowess.
Intelligence officials, including the director of the FBI, have said that they expect an increase in foreign meddling in 2020. Naturally, then, so does the Democratic National Committee (DNC). In order to emphasize the seriousness of this threat, earlier this month, at a “hackers” convention in Las Vegas, the DNC showed a video of their chairman Tom Perez apologizing for not being able to attend . . . but, of course, Perez said no such thing! He only appeared to have said this because of the work of artificial-intelligence experts, who, using a kind of photoshop for videos, made what’s known as a “deepfake.”
When done well, deepfakes are uncanny. Recently, for instance, a YouTuber used material from a 2008 interview with David Letterman to show the comedian Bill Hader seamlessly morph into Tom Cruise, then Seth Rogan. This benign celebrity deepfake went viral — deservedly, since it was well done. But what if this technology were to be used for malicious purposes? Like pornography or propaganda?
These are not academic concerns. “Deepfake” was coined on Reddit, a U. S. discussion website, by an account with the same name. The owner of this account was producing sex videos of women celebrities — fake, of course — by plastering their faces onto the bodies of porn stars. The harm of this can hardly be overstated. Even if a person knows they are watching a deepfake, some things simply cannot be unseen. Though this account (and similar ones) were removed from the site: the technological cat — growling, purring, and ravenously hungry for more — was out the bag.
The trouble is that the software used to create deepfakes is fairly easy to obtain and access.
