Deepfakes

I'm astonished that the world isn't doing something to prevent (or at least manage) deepfakes. Currently, technology can distinguish between real videos and deepfakes, but that won't last long (perhaps a few years). At that time, there will be a video showing something extremely incriminating. The defense will just say the video is fake and put the burden of proof on the world to prove it is real. The defense will than create dozens of fake versions of the video showing different people involved. Without a way of discerning which is real, will this be the end of video evidence (voice manipulation will be just easy in the near future making audio evidence irrelevant). 

To avoid (or at least delay) this problem, there must be a technological standard that functional designates video as unmodified. Furthermore, modified videos have to make it clear that it is modified (failure to do so must result in significant fines or even imprisonment). This also must include the severest penalties for any government or police agency modifying videos. 

It will not be long that on the eve of an election, there will be a video that shows one of the candidates doing something terrible. It might be proven fake, but it will be too late. 

Comments

Popular posts from this blog

Free Shakespeare Tickets - Contest #3

The Constitution and Justice

Lessons from History - Democracy