Reposting images you like or show is something many netizens have been doing. But if the image they are reposting is real or fake, how do I figure it out? It is this issue that has put Smita Sabharwal, a senior IAS official, in trouble. IAS officials reposted an image on March 31, which contained Earthmoving Machines, two deer and a peacock, originally shared by a “X” on another social media. What can netizens do to identify forgery and avoid spreading them? “There are tools available, but it’s hard for an average citizen to verify using these tools,” said Naved Khan, CEO of the Cyberabad Security Council (SCSC).
The best solution, he said, would propose a policy that requires any AI-generated content to be marked with certain identities or disclaimers. He suggested: “It’s like using TMs for trademarks or R in R circles for registration or similar AI-generated content to clearly identify them.”
How to discover manipulated images
When we discuss fake or AI-generated images, there are several ways you can discover them.
1. The AI image may have distortion, where the person may not be able to merge perfectly with the background.
2. Images may have unnatural expressions, features, shadows, or blends with backgrounds.
3. In some cases, the actual image of a person can be abused and some object is added to the background. In this case, the image on the image is searched in reverse.
4. Metadata, if any, can be checked for creation date, camera information or the software used.
5. Watermarks can be checked, as some AI tools may include watermarks.
