Deepfake Reality

In a previous post I discussed deepfakes from a hermeneutic perspective, by exploring how “a deepfake mediates how we perceive beings in the world by affording an interpretation of the fake as the real”. There is plenty of attention to the dangers of this effect and I briefly discussed its potential threats to democratic discourse. But last week we saw, to my knowledge, one of the first concrete manifestations of this threat, as European MPs were targeted by deepfake video calls of someone imitating Leonid Volkov, a Russian opposition figure. I don’t know how exactly, but apparently the culprits managed to successfully plan calls with several European members of parliament, posing as Volkov. Perhaps this is a great example of social hacking. The success of the trick can be partly explained by people being used to lower resolution video-calls during the pandemic, and I wonder whether this would have happened without our social fabric being so dependent on digital communications right now. Nevertheless, these awkward events show how deepfakes can explicitly be used to disrupt the type of transparent communication that is needed for a well-functioning democracy.

But if deepfakes can truly make the “fake” appear as the “real”, then what consequences does this have for our conception of the real?

I hypothesized that deepfakes would also have the converse hermeneutic effect of injecting “fakeness” into the real, of making the “real” appear more as “fake”. Sure, sure, that’s a philosopher’s theoretical ramblings. But in the same week of the “Volkov-incident”, something remarkable happened that exemplifies what I was getting at. A Dutch actor posted a video that was recorded by an unknown woman in order to ask him out on a date. He posted this video online because, surprisingly, a deep fake of former Dutch PM Mark Rutte appeared in this video, encouraging this actor to take her up on the offer. Dutch newspaper AD reported a subsequent backlash that prompted the actor to take down the video, with him admitting that deepfakes are “defamatory” and besides, “reality is already crazy enough”. (He could have easily added that it’s simply a dick move to post someone’s date request online. Call me old fashioned.) But another plot twist was added when research of another newspaper, NRC, showed that in fact the video was not a deepfake after all, but actually did feature Rutte. Here, the real was interpreted as fake. Since then, the original AD article has been updated and corrected.

This example is quite innocent, but nevertheless it shows how the subversive technology of deepfakes starts meddling with the interpretative processes by which we take something to be real. Perhaps the bigger picture here is not just simply that deepfakes make the fake seem real, but that they insert themselves in an ongoing historical development where the “real” loses its traditional philosophical vigor. The deepfake disturbs our comfortable conception of an image as a copy of reality (where reality is upheld as the original), instead confronting us with a simulacrum in the sense of Baudrillard: a copy without original. The deepfake gives us access to a reality that does not “exist”; and yet the deepfake obviously has a reality of its own and produces real effects. And one of these effects, I propose, is the hermeneutic subversion by which the real is reinterpreted in terms of the simulacrum. All quite postmodern, really. And granted, all this will only gain in significance the more our access to reality is mediated by (digital) imagery and the internet.

If you encounter more examples relevant to this theme, please let me know!