Tuesday, July 09, 2024

Deepfakes on the rise: How to spot one and keep secure


ByDr. Tim Sandle
July 8, 2024


Using computers to generate deepfake images. Image by Tim Sandle

Fake and misleading images regularly hit the news cycle, including attempts to trick the public with pictures of such celebrities as Katy Perry and Rihanna.

Deepfake images continue to fool millions of social media users and even news corporations. As the U.S. gears up for its election season, deepfake images are especially likely to spring up all over the place with potentially damaging consequences.

The digital identity security specialists, ID Crypt Global, have produced a guide on how people can all learn to spot fake and misleading images when we see them online. They have provided a preview of their advice to Digital Journal.

ID Crypt discovered that more than 50 percent of people do not know how to spot a fake image online. In terms of advice, the company suggests:

Location – Was an image actually taken where it says it was?

Look for landmarks, city features, and the language of signs or road markings in the background. This technique has been used to reveal faked images from global conflicts where images from previous wars are reported to be from current wars.

Shadows and light sources

You can often spot a photoshopped image through checking whether the shadows align with the light source of the image. If shadows are missing, or incorrectly positioned, the image has probably been faked.

Common sense

Common sense can help you identify a lot of fake images. For example, an aeroplane pilot apparently taking a selfie midair. An open window at this altitude would result in disaster, so it cannot possibly be real.

Hands, hair, and skin

AI images are quite easy to spot if you examine the hair, hands, and skin of the people depicted because these are things that AI really struggles with.

Reverse image search

A fake image can easily be revealed as fake by running it through Google’s powerful reverse image search engine which will usually bring up the original and true image.

Authentic Protection Media

All of these techniques are useful for spotting fake images, but they’re not always easy and not always feasible, especially if you’re viewing on a mobile device. Therefore it’s essential that news and picture agencies use technological innovations to ensure their images are properly verified and cannot be manipulated.

CEO and Founder of ID Crypt Global, Lauren Wilson-Smith, explains the usefulness of the above approach: “The explosion of fake images makes it difficult to trust what we see online. This is unacceptable and has the potential to cause great harm across the world. There are ways for internet users to sniff out fake images when they see them, but none of them are a silver bullet that’s 100 percent foolproof.”

Wilson-Smith adds: “More importantly, though, the burden of spotting fake images and disinformation should not be placed on the shoulders of the public. It is down to those who publish photographs and videos to do everything in their power to ensure that they cannot be falsified or manipulated. This is no easy task given the remarkable pace of deepfake and AI technologies, but the good guys are working just as hard as the bad to ensure that the tools are available to news agencies to protect the world from this increasingly concerning issue.”

No comments:

Post a Comment