It is quite difficult to believe anything on the Internet these days. Is the content written on the net written by the author themselves? Is the video that we see authentic?
Recently, Ranveer Singh’s Deepfake video went viral in India. In that Ranveeer Singh could be seen endorsing a political party -which is absolutely not true. The visuals of the Deepfake video are supposedly genuine but the audio is AI – synthesized clone of Ranveer. Deepfakes are shaking normal and influential people uniformly. Many other Indian celebrities like Rashmika Mandanna, Aamir Khan and many have been caught on the wrong foot for these reasons.
Video editing, picture editing are nothing new. They have been in existence since the days of Adobe Photoshop. They are now undergoing phenomenal transformations thanks to AI enabled technology. Earlier we needed Photoshop skilled professionals to do the editing, but now the digital mischief, can be done by anybody!
Here are few questions and answers about Deepfakes;
- What is a Deepfake? Deepfake is an AI generated and modified video or picture. AI and deep learning techniques learn the original content and modify it to create new content. Face swapping is one main side effect of these Deepfakes. New faces appear on other bodies and speak and say anything which are totally unrealistic. Unfortunately, it can make the modified content look absolutely real and authentic to fool users.
- Is it easy to make Deepfake content?
Yes…it is…just search for ‘Deepfake’ on Google and you get numerous tools to create Deepfake and AI generated content. Some apps to create Deepfakes are Zao, Jiggy, Wombo, LensaAI(don’t try them! :))
3. Is it easy to spot Deepfake content?
Definitely not easy…but if you are trained to spot it…you may spot Deepfake content. You do have to look at the inconsistencies in blinking of the eyes, audio, video and the lighting and the shadows of the generated content to detect Deepfakes.
And yes, there is still a possibility that you may miss a Deepfake content and think of it as the original content.
There are many tools to detect Deepfake content like Sentinel, FakeCatcher by Intel, WeVerify and more. But I am sure more tools will become available to separate the Deepfake from the original content in the coming years.
4. What will be the future of Deepfake content?
I am sure more tools will be available to create Deepfake content and more tools will be available to detect it as well.
In the mean time, we just need to keep our guard up and not trust anything on the Internet!
World is a scuffle between good and evil in multiple rounds. Good to win to keep the harmony. Good read on deep fakes. Keep going!
The more we go online the more chances of deepfake it seems.
Yes.. the times we live in… 🙂