Ever saw a picture or video online and felt that it was a fake one. Possibly it was. Deepfake in simple words is used to describe the recreation of human faces or voices easily with the help of AI. They use it to combine and superimpose existing images and videos onto source images or videos using a machine learning technique.

The term Deepfake came into existence in 2017 from a Reddit thread. With a plethora of apps and online tools available, anyone can make a deepfake. AI makes the job of recreating someone’s face so easy and fast that a novice can do it as well that too in an hour.

Undoubtedly, with all the data fed to it, AI is getting better every day. Deepfakes use AI technologies to mix pictures and create a new fake one. With time the AI learns itself and finds out the best way to superimpose pictures on one another and it to make it look as realistic as it can be. AI uses trial an error to learn how to do this task in the least amount of time, giving out the best results.

With deepfakes getting better, they possess a real-world problem on how will we be able to identify the content we are watching is real or not.

Also read: What is Formjacking, how it works and protective measures

How can Deepfakes be a problem?

Most deepfakes on the Internet are memes, public service announcements, and celebrity porn. Till now these things have been relatively easy to identify, but with advanced AI, it is becoming much harder now. These things are harmless but have proven to spread hate and anger among people.

Last year, in 2018, Rana Ayyub, an India-based investigative journalist, became a victim of a deepfake, which used her face to create fake pornographic content.

Deepfake is also used for the creation of revenge pornography. People try to make use of deepfake to make pornographic content of their ex’s, celebrities, friends and random people getting their picture from social media.

Politicians worldwide are also one of the most affected by deepfakes. Videos are altered and circulated to spread propaganda or hate.

Also read: How secure are your photos on social media? And how to protect them

Spotting Deepfakes

We can spot fake images and videos easily with naked eyes. We can spot mistakes in the pictures or videos just with a sight.

But it will not be possible for us to spot deepfakes for much longer. With every passing day, deepfakes are getting more convincing and simple to make with increasingly more sophisticated tools available online. With the help of deepfake software/apps, deepfakes can be created with a single picture.

There are multiple ways we can identify deepfake content such as AI, algorithms and blockchain technology. AI can identify deepfake remarks in a picture or a video and blockchain can verify if the media content was ever opened in a deepfake app or similar software. But till then we need to be careful whenever watching a video or picture and stop falling for Deepfakes.

Also read: What is a Whaling Cyberattack? How is it different from Phishing?