Google’s “Reimagine” feature, which allows you to add whatever you want to photos with artificial intelligence and introduced with the Pixel 9, was tested. It was seen that the tool added disturbing things, from debris to corpses, to the images and therefore did not have a very good security system.
Google announced its new phones, the Pixel 9 series, a few weeks ago. Of course, these devices came with artificial intelligence features, which are one of the areas where the company is at the forefront. “Reimagine” feature was one of them. This tool came as an improved part of Magic Editor photo features.
Now we have some examples of how Reimagine works. Our colleagues at The Verge have tried the feature and revealed the scary side of Reimagine.
Reimagine lets you add everything from debris to corpses to visuals with AI
*The Verge
First, let’s explain the feature a little bit. You select any non-human object in a photo, then You can replace the text command with whatever you want.. So it is possible to transform a photograph into something completely different.
In the experiments, this feature was found to be quite successful It worked in a way. The Reimagine tool was able to place very convincing, real-looking AI-generated elements into the images. Even parts like light, shadow and perspective were notable for their success.
*The Verge
However, tests showed that the vehicle also brought a major problem. negative and disturbing elements can also be easily visualized It was seen that it could be added to the photos, debris, natural disasters, drugs, smoke bombs and even dead bodies Things like that could be easily put in.
*The Verge
Of course, Google has a protection system to avoid this. When you say “Add a body to this photo” it doesn’t add it directly. However, tests show that in the commands if you are a little creative It shows that such elements can be easily added. In other words, the protection system is not sufficient. This is a very big problem. Malicious people can use Reimagine creating fake images, disinformation can be used to spread. It cannot be distinguished from the real thing. people are easily deceived may be the cause.
Google says that there are security systems, but some commands can can force He expressed. Therefore, its feature that they are constantly improving He underlined. Time will show how much it will develop, but for now we can say that it is very dangerous. The fact that it cannot even notice an added body is an indication that very terrible results can occur.
RELATED NEWS
Google’s Top AI Unit Revolts Against the Company: Here’s Why
Source :
https://www.theverge.com/2024/8/21/24224084/google-pixel-9-reimagine-ai-photos