It is very difficult to distinguish between fake news and real news and it isn’t just the content of the news article and the viral effect it has on social media but also the fact that it is much easier to create fake videos and photos that look real and attractive now more than ever before. Also, the act of publishing fake news is a practice that is very damaging since it affects how people behave, results of elections, and scams as well. Think of what would happen if people could create fake video clips of people that look real using just a single photo of them. Well, a new AI has been developed by Samsung and it can do just that.
Inside an AI lab located in Russia, the AI developed by Samsung was able to create a fake video clip of people using just a single photo or even a painting. Even more, the video looks pretty realistic. The AI even made a video of Leonardo Da Vinci’s Mona Lisa talking and moving, and the painting was the only image of the Mona Lisa that the AI used to make the video since it is the only one existing! Before now, creating deep fake videos required many images to be able to provide a result close to realistic. However, with the new AI developed by Samsung, you only need one photo.
Samsung named the AI “realistic neural talking heads” with the word neural being more than just a stylish choice. Neural networks are a type of machine learning capable of mimicking the human brain, and we can see that they share quite some similarities with Samsung’s new AI.
To be able to create a video clip, Samsung’s algorithm watches a lot of videos. It does this to learn how human faces move. It then applies all it has learned from the video into a single photo.
