I just came across this video of some computer scientists synthesizing Obama talking. They have trained a computer to generate a video of him saying something with a different video.
The implications of this are that, with some post processing, a world leader or really anyone can be in a video saying things that they didn't say. Therefore, no video can be trusted as proof of someone saying something.
It's very interesting but also terrifying in my opinion.
Technical article: http://spectrum.ieee.org/te...