In the interview taken by CNBC this Friday, a deepfake pioneer had stated that digitally manipulated video, which would provide a “perfectly real” feeling to all its viewers is only twelve months away from hitting the virtual market.
The University of Southern California’s computer science assistant professor, Hao Li, had made a statement that most of the people can identify nearly all the deepfakes through their naked eye only. It is very easy to determine them.
However, he has also stated that there are also few examples of deepfakes which would leave people astonished through their powerful and convincing digital representation.
He has also pointed out that there would come a time when nobody would be able to detect deepfakes anymore in any way and people would have to look for other kinds of solutions for this purpose.
A deepfake of Vladimir Putin, President of Russia has been created by Hao Liand that was showcased this week at a tech conference of MIT.
Please Also Read: Arrow Unveils World’s First Smart Shirts
Li had stated that he had created this deepfake to portray the deepfake technology’s current state. The video shows how this technology has been growing and developing at a lightning speed which is very much a shock to most of the people.
He has informed the MIT Technology Review during that time, that virtually undetectable and perfect deepfake videos are just a couple of years away.
An app for it
After this week’s interview, CNBC had asked for a clarification on Li’s timeline through an email, Li told them that the recent developments, which also include the emergence of a leading and the most populous Chinese app called Zao, has helped him to calibrate his timeline again.
Zao allows their users to swap their faces with their friends and other people. The application uses the photographs which are taken by the users. Then after that the application will then digitally insert the pictures into the scenes from popular TV shows and movies.
Even though the application has raised alarms through the world, it is still regarded as one of the most popular applications used all over China.
Through an email, he had told CNBC that everyone knows how to create deepfake videos in some ways or the other as it only requires some training backed up by data which can be implemented.
The need for more research
With the rapid advancement of artificial intelligence, deepfakes has been becoming more than believable to everyone, including the experts. It has already become difficult for the professionals to decipher and differentiate the doctored videos from the real ones.
This technological advancement has a huge drawback as now people can easily spread misinformation at the press of a button and that too within just a few minutes. With the presidential election coming in 2020, this has raised danger signs to an utmost level.
There is an urgent of new researches who would bring new, effective, and efficient solutions to help in determining the deepfakes from that of the originals.