Deep Fakes: Deception, Mimicry, and Bullshit

A deepfake is a specific type of media where a person in a picture or video is replaced with another person's likeness in the hopes of being undetectable. Though they were initially introduced in 1997 as a way to make a person look as if they were saying something they weren't, the term "deepfake" was developed very recently on Reddit as celebrities' faces were swapped onto the bodies of actresses in pornographic videos.  

The popularity of deepfakes have grown as the necessary technology and artificial intelligence mechanics have progressed to a point that we can no longer confidently identify or distinguish a deepfake from a real video. In one of my favorite examples, Bill Hader, a popular comedian, does his impressions of actors Al Pacino and Arnold Schwarzenegger. Look closely, and you'll see that while Bill Hader impersonates these other actors' voices and mannerisms, his face also seamlessly transitions to look more like the actor he is impersonating. The software used for editing this video, DeepFaceLab, is readily available on GitHub. The question to be asked, however, is how dangerous can this technology be when placed in the wrong hands? 

Harry Frankfurt, an American philosopher, suggested that tools like these, when given to those who are willing to lie and deceive, only contrive to interfere with and impair our natural efforts to apprehend the real state of affairs.

Donald Trump's face pasted onto actor Alec Baldwin's body

With the potential of being weaponized, CNN reported on "The Pentagon's Race Against Deepfakes". Politically, the ability to efficiently and effectively produce convincing misinformation with an underlying purpose can undeniably be utilized as a weapon. Finding a person who looks and sounds similar to a prominent social or political figure is not difficult; recording a video in which you imitate this figure in a negative light and run it through the software is now even easier with deepfakes. Thus, social media apps like Facebook are forced to ban deepfakes prior to the 2020 US elections and are now heavily investing in software that would attempt to distinguish between a real video and a deepfake.

The intention behind creating deepfakes have (so far) never been done with malicious intent. But as the technology gets stronger and the artificial intelligence gets smarter, there will no doubt be those willing to make use of the deception with various motives. As Frankfurt explains, insofar as this deception succeeds, we acquire a view of the world that has its source in the imagination of the deceiver rather than being directly and reliably grounded in facts. Although today this technology is only used by the public for fun, as it evolves we must be wary of what's real and what isn't.