The Rise of Deepfakes

6 months ago by Sophie Stones

The Rise of Deepfakes


Over the weekend a new app has been blowing up in China. Chinese face-swapping application Zao, a subsidiary of Momo Inc. soared straight to the top of the Chinese App Store chart due to its easy to use viral format. Swapping the faces of its users with celebrities and films stars in hundreds of video clips, Zao is the latest in deepfake technology, allowing ordinary people to utilise content manipulation which used to take real skill, by simply uploading one photograph.

Like FaceApp and other photo software which came before it, Zao has courted controversy already through their ‘terrifying’ terms of service, which state they have “free irrevocable, permanent, transferable, and re-licensable rights” to all of the information shared with them. (EDIT: After their ratings plummeted due to angry users, Momo Inc. have since revised the release to state that they will only use pictures or mini-videos to improve the app) However, with this new deepfake technology has come renewed fears that in a very short space of time, we may be unable to distinguish completely between what is real and what is not.

Actor, Comedian and Filmmaker, Jordan Peele, recently implored viewers to question what they see, all through a manipulated video of Barack Obama, in which the former President called Trump “a total and complete dipsh*t”. Whilst Peele made it clear that his video was fake, it was physical proof that others may not be so honest or fair. In fact, deep fake videos are already being used for nefarious reasons, with women’s lives being torn apart as their faces are placed into x rated videos.

In another instance of misuse earlier this year, Trump supporters attempted to discredit Nancy Pelosi by posting a video in which the Democratic House Speaker appeared to be drunk and slurring during an interview. Whilst the video was found to be doctored, Facebook refused to remove it from their platform, heightening concerns of the spread of misinformation.

The Pelosi video, which used relatively low tech compared to Zao, by simply slowing down her speech, is a significant example of deepfakes past. With images of migrants being used within the wrong context to cause uproar by Nigel Farage and other far-right politicians, people haven’t needed GAN to fool people in the past. This, Claire Wardle, an online manipulation expert, says is why we all just need to calm down when talking about these new technologies.

Stating that “the alarmist hype is possibly more dangerous than the technology alone”, Wardle is less concerned that people will be duped by fake videos, which are often if not always exposed, as she is that “it becomes much easier for the guilty to dismiss the truth as fake”. Wardle’s fears recently came to fruition, as Donald Trump, who had previously apologized for comments about grabbing women, backtracked, now saying he doesn’t think he actually said it.

You’d be hard pushed to find anybody who hasn’t heard the term ‘fake news’ but with the rise of deepfakes this is becoming less of a slander on journalists and more of a reality. It’s now up to both social networking sites to monitor information being put onto their platforms, and ourselves, who need to question; does this seem right?