The video above shows a still-in-development system called Face2Face (research paper here) created by researchers at Stanford, the Max Planck Institute and the University of Erlangen-Nuremberg. It would allow you to take YouTube video of anyone speaking, and to pair it with a standard webcam video of someone else emoting while saying something entirely different. Thehe Face2Face system will synthesize a new video showing the originals speaker making the second speaker’s facial movements, including the interior of the mouth, so it looks like the original speaker is saying what the second speaker was.
Tech Crunch reports that the system isn’t quite ready for market yet. Gee, I can hardly wait. This “advance” has the potential of making video just as unreliable and untrustworthy as still photography is now. Web hoaxers, Ted Cruz’s marketing team, unscrupulous political websites like Breitbart and others will have a field day once Face2Face is perfected.
The justification for creating such technology is the same as the rationalizations behind cloning velociraptors in “Jurassic Park”: because we can, and because we can make money with it. Can any good come from Face2Face? It’s late and I’m not at my best, but it seems to me that the end results of having another tool for liars just means more lies, more cynicism, more misinformed people, and less trust.
Isn’t it irresponsible and inherently unethical to invent something like this?