FALSE WITNESSES Fact and Fiction in the Age of Fake News

‘President Trump is a total and complete dipshit,’ says former US president Barack Obama in a YouTube video. Except it’s not really him talking; instead, this is a fairly convincing digital creation produced by BuzzFeed, with writer/director Jordan Peele lending his voice to doctored vision of Obama speaking. The video purposefully draws our attention to the rapidly improving technology that can be used to create fake news. ‘We’re entering an era in which our enemies can make it look like anyone is saying anything at any point in time, even if they would never say those things,’ Peele (as Obama) says in the video. ‘Stay woke, bitches.’

The technology to create and spread fake news is developing at an incredible rate. BuzzFeed’s video demonstrates the ease with which software can be used to achieve convincing digital face swaps. Known as ‘deepfakes’, these videos are the products of machine-learning technology that references photos to replace a face frame by frame. People have used the technology to insert Nicolas Cage into movies he never starred in, made Alec Baldwin’s Saturday Night Live impersonation of Donald Trump more realistic and plastered the face of Elon Musk onto Robert Downey Jr’s Iron Man.

While the results are not always entirely convincing, they have all been achieved using off-the-shelf computers, and the technology is improving rapidly. To demonstrate the power of this software, users have shown how they can replicate expensive visual effects employed in Hollywood films. In 2017, while shooting Justice League (Zack Snyder, 2017), Henry Cavill was contractually prevented from shaving the moustache that he had grown for Mission: Impossible – Fallout (Christopher McQuarrie, 2018). As a result, Warner Bros. spent a considerable amount of time and money on post-production to remove the moustache from Cavill’s Superman, with less than impressive results. In response, a programmer re-created the effect using a secondhand computer.

The software used to create convincing fakes doesn’t stop with video. In 2016, Adobe announced Project VoCo, which can create a digital simulation of someone’s voice with little more than a twenty-minute audio recording of them speaking. Another company, Lyrebird, currently has software