Counterfeit content round the corner
By Teamspirit on Tuesday, 11 July 2017
Developments in machine learning, a type of artificial intelligence (AI) technology, could soon mean that computer-generated forgeries of original content are appearing more frequently in your feeds.
The manipulation of imagery and videos is nothing new - humans have been adapting these media from the outset for the sake of art, satire and also fraud.
What makes the introduction of this new generative adversarial network (GAN) technology into the mix so significant is that it requires no trained skill - and most importantly since the content is generated by a computer, other computers will regard it as real.
Unlike fake copies created by humans, which may include an incorrect pixel or two, this method of automatic content creation is required to have fooled a piece of verification software in order to have been generated in the first instance.
At the moment this method of content generation is only possible at low resolution, but as the software becomes more powerful it could well be that we are soon witnessing fake content which is harder and harder to decipher from the real thing.