Disguising Data With Inversible GAN
There is a method to disguise data not by disguising them as HTTPS traffic which is unreadable, but simple HTTP as well as other protocols for diversity through AI. AI is mostly used on the censor's side as to distinguish tor traffic from normal traffic, on the other hand, a GAN is made just for that. A GAN fools not just the human eye, but rather an AI as well. This means if one could invert the generator to map the outputted image into the source image, they could "encrypt" the data into the "image", or rather something like a HTML file, and "decrypt" it with the inverse. Here is a prototype for the inversion https://github.com/Aphrodites1995/inversegan.
Trac:
Username: Aphrodites1995