Real-time Deepfake AI is getting so advanced that you can no more differentiate between what’s real and what’s fake.
Most social media have strict regulations to reduce the circulation of these Deepfakes. But they are evolving at a rate with which it is close to impossible to find them all.
The common public should have a basic knowledge about these in order to avoid getting tricked or fooled. There is one such incident which happened recently.
Check this out:
Ali Aliev’s Prank
Pranks done on common people shouldn’t get supported. But there is no denying that this prank is cute and heart-warming.
Ali Aliev is a programmer who has developed a method to use real-time Deepfakes in video conferences.
To test it, he joins in a random meeting in Zoom. This video is now so popular on the internet.
It is probably one of the cutest zoom bombings ever. The people in the meeting did not have a single hint that it was fake. The Deepfake was so accurate that it was hard even to doubt.
Why Was It Hard To Doubt?
You can easily identify a Deep Fake when it is a picture. There are various ML algorithms that can separate a Deepfake from the real one.
But in this case. Ali Aliev uses a Real-time Deepfake AI. He also uses it on a live video conferencing platform.
Online video calling platforms have a major drawback. In order to remove any lags in the meeting, the frame rate of the video gets reduced.
This reduces the overall video quality. But on the upside, you get to experience calls without any latency.
This provides a convenient environment for Real-time Deepfake AI to operate. As the video is not clear enough, the job of identifying the AI becomes difficult.
Avatarify
Avantarify is the technology which Ali Aliev used to make it possible. This is something which he developed himself.
He says that he used “The First Order Motion Model For Image Animation” to make this possible.
It is nothing but an open-source AI code developed by the brilliant minds at the University of Trento, Italy.
It is available on Github for anyone to try and experiment with it. Ali Aliev seems to have mastered it.
Requirements
The Deep Fakes we know can function only using pre-recorded video files. But Avatarify is different.
It can process videos in real-time. But requires a PC with top-notch processors and RAM. You need a lot of processing power to make this possible. You also need to train the AI for it to work seamlessly.
But the developer says that you can do it simply by putting all of the person’s images in a folder. Needless to say that you will need a keen knowledge of programming and AI.
Conclusion
No matter how interesting the technology seems to be, it always comes with a lot of risks. Gradually people would start using it to trick and steal from people.
It is important for users to know the consequences of using Deep Fakes. Various countries around the world have strict laws against AI users. Make sure you never get into trouble.
