Fake new world

After images and text, the Hnal frontier was video and voice. An impressive technology called Deep Fakes managed to overcome this frontier. Given a video and pictures of a person, it is capable of swapping an actor’s face with that of the person


A few decades ago, before the advent of the World Wide Web (WWW), chatrooms were a popular pass time. In those days, people realised (at their own expense) that the person on the other side of the screen might not be the person who he or she claims to be.

There were cases where people started a virtual relationship which eventually blossomed in an engagement. But others were not so lucky, and in a particular case, a woman in the US even tried to kidnap (in real-life) her ex-virtual boyfriend.

These stories did not end with the arrival of the WWW, but they took over a whole new level. People are now portraying a perfect life on social media networks, and some of them are the envy of the digital world, living a life of champagne and jet-setting. Of course, most of it is fake.

Worse than that, some people even go the extra mile and create fake profiles. In some cases, these people are harmless. For them, it is merely an online game, an escape from their real-life. But there are cases where these people are predators, preying on the most vulnerable of people.

Some of these are even paedophiles! We’ve seen various instances reported in the media where children make friends with another person, thinking he or she is a peer, only to discover when they meet them in real-life that the person is a paedophile. Unfortunately, some of these stories do not end well.

However, if you thought that this was the maximum of online fakeness, you’re in for a shock. The fake world we’ll experience in the coming years is way more advanced. With the rise of Artificial Intelligence (AI), a new area of research has emerged known as Synthetic Media (SM). The idea behind SM is that an AI can be used to generate images, text, voice, videos and practically any other media.

Manipulated images have been around for years. Photoshop (one of the most eNective programs used to edit pictures) is capable of modifying existing scenes or even creating fake ones.

However, just last year, Nividia announced GauGAN, an impressive AI system capable of converting children like scribbles into high-resolution photos. A simple line with a circle changes into a realistic picture of a waterfall next to a lake. Of course, the proliferation of manipulated photos will likely increase. Expect to see more pictures of friends travelling to fake exotic locations, showcasing their digitally enhanced bodies, while living their depressing real lives.

Interest in automatic generation of textual-content started in the 50s, but in recent years, there have been some signiHcant breakthroughs.

  • A computer program called “Eugene Goostman” (which gave the impression that he was a teenager) was very close to fooling experts who were chatting with him, in thinking that he is a human being.
  • An AI system wrote a novel which managed to pass through the Hrst round of a Japanese literary prize.
  • Microsoft created a Poet AI capable of creating poems, some of which even fooled human judges.
  • The Press Association in the UK is making use of AI to select, adjust and publish news stories.
  • And with 300,000 chatbots currently live on Facebook, it is very likely that when you lodge an enquiry through a Facebook chat, you will meet one of them!

But this technology has its negatives as well. OpenAI has just released an AI fake text generator. Its creators claim that the system is so good at creating fake news that they labelled it dangerous. So expect a new wave of fake news and mass-generated spam to hit us in the coming years.

After images and text, the Hnal frontier was video and voice. An impressive technology called Deep Fakes managed to overcome this frontier. Given a video and pictures of a person, it is capable of swapping an actor’s face with that of the person.

This technology is revolutionary for the movie industry since it allows Hlmmakers to make movies at a fraction of the cost. Unknown actors can act in a blockbuster Hlm only to have their face swapped with a celebrity’s face in the last minute.

However, this tool can also have some dangerous uses. Deep Fakes replaced the face of a pornstar in a porn movie with that of the famous actress Gal Gadot (Wonder Woman). In another video released online, the former president of the United States Barack Obama slams Donald Trump. And the list of similar cases can go forever. Of course, they are all fake videos.

In the coming years, expect more of these manipulations. However, the AI algorithms are becoming so good that it will be incredibly hard, if not impossible, to distinguish between truth and fiction. In these cases, there is only one hope for us, the rise of new AI systems capable of spotting these manipulations, alerting us and stirring us away from this fake new-world.

More in People