“Deep fake” video will be able to show people saying, with the authentic ring of their own voices, things they never said. It will show them doing things they never did, by melding their images with other video or creating new images of them from scratch.
At a political level, deftly constructed video could show a political leader advocating for the reverse of what she stands for, or portray bloody events that never happened. It could trigger riots, swing elections, and sow panic and despair.
At a business and personal level, it could be equally dangerous. Fake statements by chief executives or banking officials could throw financial markets into turmoil. False videos could be created about anyone’s private life, with devastating effects.
In some ways, deep fake videos are no worse than rumors or false news stories. People will believe and share anything they find engaging or that reinforces their beliefs.
But videos are more dangerous because of the authority this medium has taken on in society. For years, video has been the ultimate argument-settler. Online news outlets routinely hyperlink videos into stories to buttress the credibility of their reporting. Dash-cam video is often the clincher in claims of police malfeasance.
Society now has to learn that video no longer guarantees reliability. Instead, it could be the biggest lie of all.
Faced with this menace, experts have looked to search engines and social media platforms to identify and expunge deep fakes, and to governments to find ways to outlaw them. In an excellent analysis published last month on the Social Science Research Network, law professors Robert Chesney and Danielle Citron reviewed some possible tactics, and their downsides.
Given the existential danger to truth that deep fakes present, the media industry must make its own efforts to counter the threat — and it can be effective, if it acts quickly.
Right now, we have a window of opportunity in which most fake videos are far from perfect. In many cases, it’s laughably obvious that the images are manipulated. In others, fake video looks convincing to the naked eye, but technical tests can easily demonstrate that it’s a fraud. But the quality of technology we already see in the best science-fiction and fantasy movies, as well as the most sophisticated video games, shows us what will inevitably be available for broader exploitation — to far more ominous ends.
This window gives us a chance to sound public alarms about deep fakes, promote basic tips on detecting them and inoculate audiences against the better fakes to come. It’s not an impossible lift; the public is well aware of fabricated pictures and fake news stories (even if, lamentably, they still share them). News coverage of the dangers of deep fakes will find a ready audience.
News companies can also publicize analytics sites and browser extensions that can alert viewers to video fakery.
This is not only a concern for traditional major media but also for specialized outlets. Fake video could show football teams scoring touchdowns that never happened, or a singer’s voice cracking when in fact she sang perfectly. Media specializing in all subjects should be sounding warnings.
Unless the dangers of fake video receive broad public attention now, the public will be caught unaware when truly convincing fakes appear, perhaps with disastrous results.
Finally, in publicizing the dangers, media need to avoid a tone of hopelessness — “Soon we may never know what is real and what isn’t.” Quality media outlets need to emphasize how carefully they vet video. They should make sure their ethics codes and verification procedures adequately address the dangers. Otherwise, audiences will doubt any video — including legitimate and important footage that media outlets gather in their own breaking-news coverage and investigative work.