There sits Mark Zuckerberg, staring into a camera, proclaiming that the secret to controlling the future is controlling the stolen data of billions. A shocking declaration for the head of one of the world's most important and powerful companies to make. The only problem? Realistic as the video appears, it's fake — or, as these types of videos have come to be known, a deepfake.
The clip, shared on Instagram, came from the artists Bill Posters and Daniel Howe and was produced as part of an art exhibit called Spectre that recently premiered in the U.K. It did exactly what art should do: poked not only at Facebook's power, but also at the company's own moderation policies that state it will not take down footage that is fake. Recently, doctored footage of Nancy Pelosi, edited to make her seem slow and her speech slurred, was circulated on Facebook and the company refused to take it down, claiming that such content did not violate the company's guidelines.
Deepfakes are held up as a terrifying example of the sinister power of technology. They are the result of sophisticated technology that allows creators to make it appear that someone is saying something they never did. Last year, there was a clip that purported to show Barack Obama but was in fact the voice of comedian and director Jordan Peele, synced to Obama's face. Glanced at, it was pretty believable. Even just 18 months later, the technology has progressed at a frightening pace; now, you can simply type in what you want someone to say, and software can generate both believable video and audio of almost anyone saying almost anything.
Deepfakes are unnerving because of what they augur: a future in which it may become impossible to trust anything we see. But for all the panic, deepfakes themselves may be something of a red herring — that rather than simply being the result of a new technological problem, our current political moment is instead the product of how a broader set of material circumstances interact with the new world of digital technology.
In one sense, deepfakes aren't entirely new because image manipulation isn't new. Photo editing, camera trickery, and arguably even the simple existence of things like fiction and satire have always made us skeptical of what we see. As each new thing emerges, we gain a media literacy in response and learn to look critically at what is in front of us.
But deepfakes don't just seem to threaten our relationship to the image; they instead poke at something discomfiting about the idea of shared reality. Everything from political consensus to social cohesion depends on the idea that we are all roughly talking about the same thing. Deepfakes join the turn toward fake news and the mainstreaming of conspiracy theories which make it seem there are competing worlds in play — one in which, say, climate change and vaccine science are real, and another in which they are not. What a fake but mostly convincing video of someone speaking does is cement the sense that the facts that make up our shared sense of reality have become unstable, that instead of building a worldview out of evidence, we build it out of assumption and bias and simply seek out things that support what we already believe.
Yet, oddly, perhaps this very worrying idea is precisely why deepfakes aren't as terrifying as they initially appear — or at least, why the worry they represent isn't entirely new. We have, after all, always had an ideological relationship to reality. While we often think of the progression of history as an accumulation of facts, when we think of, say, the rise of feminism or the criticisms of colonialism, the "evidence" at their base was there all along — what was needed was to reshape how we relate to the world. Ideology has always been a method of structuring reality.
It is thus tempting to say that deepfakes are just an extension of this reality and are nothing to worry about. But this isn't quite true, either. Instead, the emergence of each new technology has particular effects on our understanding of the world. The mere existence of technology to manipulate images and video is itself enough to sow doubt along ideological lines.
Consider: In the early 2010s, news broke that the now late Toronto mayor Rob Ford was seen on video smoking crack. Almost immediately, skepticism arose because the video might be doctored, despite the fact that it was viewed by two reporters who had seen Ford up close hundreds of times. What mattered was not plausibility — that some random person could have believably created a video of a big city mayor — but instead plausible deniability: that the simple possibility that one could in theory manipulate video meant that if you were inclined to support the populist mayor, the video must have been false.
Deepfakes are thus neither some novel terror nor are they insignificant. Instead, what they highlight is the role media plays in our current politics. The division between left and right isn't disagreement about policy; it's about competing understanding of what reality actually is. And what digital technology of all kinds allows us to do is form communities around these realities — in Facebook groups, on Twitter, on Instagram, building a mediascape for ourselves composed of what we want to see and what we want to believe.
In that sense, it is the broader material effect of digital media that is the key change — not just deepfakes or Photoshop, but the breakdown of barriers to broadcasting and publishing that in turn allows "realities" to form, replete with their own communities, media, memes, and in-jokes. Deepfakes themselves may not challenge our reality, but it's because at this moment in our history — full of digital manipulation, political polarization, and a willingness to stay ensconced in one's own communities — there's not that much left to be challenged.