Deepfake images are fake images created using deep learning techniques.
The images created by the deep learning algorithms can be pretty convincing.
To date, deepfake images have been used to create revenge porn, fake news, and various hoaxes.
Today we are going to play a little game to see if you can spot a deepfake!
But first, let’s go through how they are made in the first place.
Producing false Images using deep learning
Deepfakes are created using a deep learning algorithm known as a GAN.
Two parts make up GAN or Generative Adversarial Networks that support the algorithm. The first is the generator. The generator is trained to create art based on the training data set.
The second is the discriminator.
The discriminator decides whether or not art generated by the algorithm is ‘real.’ Real means that it looks like art from the training, i.e. real art.
In successful GAN artwork, the generator fools the discriminator into believing it is real.
So you think you know what a deep fake is? Well, now it’s time to see if you can identify a deepfake.
Try the quiz, can you spot a deepfake?
What are your thoughts on the future of media and deepfakes?
So how did you do with the quiz? Let me know your score in the comments!
The images were taken from the Reddit Deepfake thread.
As you can tell, it is becoming increasingly challenging to spot deep fake images online.
It is important that we understand not everything we see online is real. First fake news and now deep fakes, we all need to be aware when online.
So long as we’re vigilant, we can ensure we are not falling prey to the fakes.
