Obscuring the Lines of Truth: A Blurred Reality with Deep Fakes in Post True World

Deep Fakes

Technology has always in the forefront and sometimes considered as a necessary evil and when the world media is struggling with fake news, the advent of Deep fake has tilted the balance more towards the evil side... So, what are these Deep Fakes?? 

Deep Fakes are an Artificial Intelligence (AI) based technology used to morph a person’s face onto another and manipulate voice recordings. Although, Computer Generated Imagery (CGI) technology has been already in use mostly for movies and advertisements but the same required expensive software and sophisticated technical skills, and were prepared in specialized laboratories. 

With the advent of new researches in Deep Fakes, I came across an interesting research paper titled "First Order Motion Model for Image Animation". The framework in the paper addresses the age old sophistication problem, without using any annotation or prior information about the specific object to animation. Once trained on a set of videos depicting objects of the same category (e.g. faces, human bodies), the method can be applied to any object of the class. 

So decided, why not give it a try myself as a test subject on the model framework described in the paper. 

You can even trigger responses to cartoons or any objects for that matter, with much ease. Here's some test results that I tried. 

The boundless capabilities of Deep Fake technology and the border-less nature of the interest has indeed blurred the lines of truth and pushed us towards a Post- Truth world. But as Steve Jobs once said. "Technology is nothing, what's important is that you have faith in people that they are basically good and smart, and if you give them the tools, they will do wonderful things with them". 

Link to the paper "First order motion model for image animation"- https://papers.nips.cc/paper/8935-first-order-motion-model-for-image-animation.pdf