The rise of the deep fakes

In one episode of the popular TV series, The Blacklist starring James Spader as Red Reddington, a scientist was confronted and shot by security operatives while trying to gain entry into his place of work, why? He had apparently been seen in a video published online threatening to carry out a bombing with an explosive which had prompted the security operatives to engage him with lethal force. The security operatives had further being convinced of the authenticity of the video as the man had arrived at his place of work holding a cellphone which they assumed was the detonator for the bomb prompting them to open fire on him.

It turned out that there was never any bomb or bomb plot. The supposed bomb threat video was a manipulated video by an advanced artificial intelligent entity with the aim of eliminating the man by presenting him to be a bomber. Raymond Reddington eventually got his hands on the technology to suit his own purposes. He specifically utilized it to expose one of his close associates who had planned to betray him to one of his many adverseries.

This episode of the series was my first encounter with a technology which is referred to as Deep fake. So what is a deep fake? According to an article in the Guardian Newspaper dated January 13, 2020 Deep fakes are images of fake events produced using a type of artificial intelligence called deep learning. The same technology has also been applied maliciously to manipulate audios and videos.

According to the Guardian article, deep fakes were birthed in 2017 when a Reddit user posted manipulated pornographic videos where the faces of popular screen actresses had been used to replace those of the persons actually involved in the video. Although many of the deep fakes are of pornographic nature targeting celebrities especially women, the use of the technology for sinister motives has been used in diverse ways such as committing fraud. In recent times deep fakes have made their way to the political arena being used to manipulate videos of political figures with malicious intent.

Videos of former U.S. Presidents Barrack Obama, Donald Trump, Speaker Nancy Pelosi and Russian President Vladamir Putin just to mention a few, have been manipulated via deep fake technology to serve the purposes of their opponents or critics. A more recent deep fake video which has received a lot of attention is one featuring present U.S. Vice President Kamala Harris.

In the deep fake video which has been manipulated she is seen to be making a speech where she states that most of the dead as well as those hospitalised due to Covid-19 were individuals who had been vaccinated against the virus. The authentic video broadcast of the VP's speech shows her saying the exact opposite of what was conveyed in the deep fake video which was indeed very convincing pointing to the frightening level of sophistication that the technology has presently attained.

Coming closer home, in the aftermath of the events of the evening of 20th January, 2020 in Lagos State Nigeria during the protest against police brutality and misconduct, a lot of information was flying about as to the casualty figures and what exactly transpired that night. One of the media that was spread via social media was a video of a military officer firing a high caliber weapon mounted on a gun truck into the dark. This video was eventually revealed to be a deep fake as the said event was reported to be one from the fronts at Sambisa in North Eastern Nigeria where the military was engaging book haram terrorists. To make the video convincing, the manipulators had inserted the chants of the slogan for the protests.

The reality is that we are likely seeing the tip of the iceberg as to the power of deep fakes as machine learning is constantly advancing in leaps and bounds. Deep fakes have the potential to be used in manpulating the outcomes of elections by swaying public opinion in favour of or against a particular candidate. It could even be used to instigate civil unrest within countries of conflicts between nation states by targeted disinformation campaigns spread via deep fake videos and audios.

While previously making a deep fake required some level of technical skill, the deep learning technology is now being pushed to mobile platforms with apps now available to replace faces in photographs. Thus the ability to make deep fakes will virtually be available to anyone with a smart phone or a personal computer which meets the requirements to run the software for this purpose.

With the way deep fake technology is evolving, the popular saying that seeing is believing may well be on its way to oblivion as the barrier between truth and falsified information is being torn to shreds.

Shukran

Danke

Merci

Gracias

ขอบคุณ

ধন্যবাদ

Спасибо

for reading.