In a digital age plagued by fake news, reality has been upended— again— by “deepfakes”. An amalgam of the words “deep learning” and “fake”, deepfakes are videos which have been edited through artificial intelligence, with a high capacity to dupe viewers. While counterfeit content itself is not new, deepfakes go beyond everyday techniques and effectively stitch the likeness of persons onto dynamic media, blurring the real and sounding off concerns regarding the abuse of such technology.
Prophetically, the late postmodernist philosopher Jean Baudrillard wrote: “The media represents a world that is more real than [the] reality that we can experience. People lose the ability to distinguish between reality and fantasy. They also begin to engage with the fantasy without realizing what it really is.” In much the same way that fake news confounds a viewership, deepfakes simulate a reality that never existed to begin with, provoking a response that is unwarranted and misdirected, while inadvertently bolstering the “realness” of the unreal.
Deepfakes dismantle any remnant of authenticity video possessed over image. As a technologically adept society, it is understood that images can be, and are, readily doctored and promulgated, informing a reluctance to outright trust images of people found online, but videos of people up until now represented a greater degree of authenticity.
Unlike the editing of a single still image of a person, matching and altering every human gesture and expression has proven a difficult feat without failing to appear off-putting and noticeably fake. Deepfakes intend to (un)remedy that, by matching and altering every human gesture and expression in such a way that it approaches authenticity.
While the technology is not quite infallible yet, the threat is the implication of a future where misinformation is indistinguishable from the truth. It is the ability to defame and deceive with the (falsified) evidence to back it up. A society fueled by misinformation is ultimately a society destabilized.
Moreover, in the same way, people have an inclination to assume images have been photoshopped, individuals can further muddle up the real and unreal by incorrectly assuming real videos are the product of deepfake technology.
Reality and fantasy will be one and the same.
However, despite the unmistakable propensity to view deepfake technology as wholly sinister and destructive to the supposed boundaries between the real and unreal, artists have done their own upending by wielding deepfake technology to create art.
On one end of the spectrum, aptly named Gillian Wearing makes uses of deepfake technology to repurpose her face as a mask that actors then wear. The video distortion effects and eerie music coupled with the shift between actors wearing Wearing’s deepfake face and the jumble of voices creates a deeply unsettling experience.
Playing off of the signified dystopian aura that deepfakes allude to, Wearing constructs a rather uncanny advertisement, laying bare the anxieties that deepfake technology engenders.
On the opposite end of the spectrum, Collider Videos harnesses deepfake in a much more lighthearted manner, for entertainment. Placing the likeness of celebrities onto skilled vocal impersonators, Collider Videos constructs a comedic rendezvous unlikely to ever happen in reality.
The interplay between deepfake’s inherent attempt to pass for reality is undercut by the glaring mismatch between the simulated faces and the bodies of the impersonators, and by the comedic subject matter. The result is a subversion of the sinister connotations surrounding deepfake technology; with laughter, Collider Videos assuages— for a brief instance— fears of a looming dystopia.
The Race Against Deepfake
As deepfake technology continues to evolve, a future iteration will spell the end of truth itself, rendering the real indistinguishable from the unreal (something Baudrillard says has long since happened).
Fortunately, counter-technology is being developed at a rate that seeks to actively outpace advances in deepfake technology. Such technology seeks to detect and distinguish deepfake videos, delineating the unreal from the real and preserving some semblance of truth.
However, detection is only half the battle. Droves of people have shown that they are willing to swallow fake news, even after it has been revealed to be fake. The other half of the battle is one against human biases, which ultimately rests on a viewer’s propensity for healthy skepticism.
Besart is a contributing writer at Glossi Mag.
He is currently a student and aspiring writer living in Los Angeles. He loves running at the beach and staying hydrated.