What you hear right now, I never said.
What you hear is an AI generated clone of my voice that was created by running a written script into an audio deepfake application called Speechify. The creation of deepfakes is now easier than ever and virtually anyone can do it.
In some instances, this synthetic media content can be entertaining and even funny, like that photo of Pope Francis in a white puffy jacket. In other cases — like the Joe Biden robo-call — they can be dangerous and even threaten the integrity of the democratic process. But the real problem is not that deepfakes can momentarily fake us out, the problem is that the existence and proliferation of this AI generated content destabilizes the assumed relationship between media representations and reality, making it increasingly difficult to decide if what is seen or heard is actually real or AI generated.
Today we take it for granted that a picture or an audio recording is an index or faithful representation of something that is really real. You hear my voice on the radio, and you immediately assume that there must be some real person who actually spoke these words. But deepfakes flip the script on us, challenging this very assumption. At this point in time, the rules of the game — at least as far as we thought we understood them — no longer apply. Reality itself has become virtual.
Welcome to the desert of the real.
I’m David Gunkel, and that’s my perspective even if I never said it.
