Deepfake has been dubbed as the 21st century's answer to Photoshopping.
A deepfake video showing actor Rashmika Mandanna is gaining traction on the internet. The clip shows Ms Mandanna entering an elevator. But fellow actors and other users on social media platforms like X (formerly Twitter) quickly pointed out that it is digitally manipulated. The video shocked actor Amitabh Bachchan who called for legal action and attracted the attention on Union minister Rajeev Chandrashekhar. The video has been manipulated in such a way that it is very difficult to make out that the character in it is not Ms Mandanna.
What are deepfakes?
It has been dubbed as the 21st century's answer to Photoshopping. Deepfakes use a form of artificial intelligence called deep learning to make images of fake event, hence the name.
According to Tech Target, deepfakes transform existing source content where one person is swapped with another - in this case, Ms Mandanna's face is swapped with British-Indian woman Zara Patel.
Such video are also used to create entirely original content where someone is shown doing of saying something they didn't do or say.
How are they made?
The first known example of widespread use of a deepfake video was spotted on Reddit in 2017, when a user swapped the faces of celebrities - Gal Gadot, Taylor Swift, Scarlett Johansson and others - to create porn clips.
These videos are create by machines that use deep learning algorithms. They scan through thousands of face shots of using an AI algorithm (called encoder) that allows the machines to learn similarities between the two faces and reduces them to their shared common features, compressing the images in the process.
The encoder then feeds the wrong image to the original source and another algorithm called the decoder reconstructs the face with the expressions and orientation.
For a convincing video, this has to be done on every frame.
What are deepfakes used for?
Most of the deepfake videos have been pornographic in nature. But during elections time, digitally altered clips of politicians are also circulated to falsely attribute a statement or promise to them.
A few years ago, former US President Barack Obama was seen calling Donald Trump a complete dipshit in a widely circulated deepfake video. Similarly, Meta chief Mark Zuckerberg was seen bragging about having "total control of billions of people's stolen data" in another such video.
AI firm Deeptrace found 15,000 deepfake videos online in September 2019, a near doubling over nine months. Experts like Danielle Citron, a professor of law at Boston University, have raised serious concern about the technology. "Deepfake technology is being weaponised against women," he told The Guardian.
Is deepfake used only to create videos?
The AI-driven feature is not just used for video, but also to create entirely fictional photos from scratch. Cyber criminals have used the technology to create the profile of "Maisy Kinsley", a non-existent Bloomberg journalist on LinkedIn. Another such fake profile was that of "Katie Jones", who claimed to work at the Centre for Strategic and International Studies.
Who is making deepfakes?
The Guardian said that everyone from academic and industrial researchers to amateur enthusiasts, visual effects studios and porn producers are using the technology to create manipulated videos.