Samsung’s Galaxy S24 can remove its watermarks which are meant to show that the photo is fake
the New Samsung Galaxy S24 It’s all AI-driven, so much so that the phone maker has made a big deal about how its new AI watermark aims to discourage users from posting AI-edited or outright fake photos online. The problem is that S24 It features so many AI capabilities that a simple Magic Eraser-like tool can remove the same watermark without any fuss.
Then Galaxy is open At an event last Wednesday, Samsung announced that any image using the company’s production capabilities was accompanied by a watermark. This appears in the lower left side of each image as the Samsung “Sparkle” logo.
Gizmodo has confirmed that Samsung’s AI-based object scanning tool — an advertised feature that can remove and replace unwanted parts of an image — can remove the watermark that is supposed to determine whether an image was created by AI. This worked on Galaxy S24 Ultra Test model, and a few others have also noticed the very significant censorship aimed at suppressing fake images from making the internet even crueler than it already is.
But it gets even more confusing. In Gizmodo’s tests, some images where we used the object erase tool didn’t appear with a watermark, while others did, including an AI-adjusted wallpaper. In Samsung Re-release of the Unpacked conference recording, the watermark appears once Mr. Beast uses the object eraser to get his crane off of the giant (bad-tasting) chocolate bar. In the text at the bottom of the same clip, Samsung made clear that it doesn’t take the blame for any inaccurate deepfakes.
“General editing requires a network connection and Samsung Account login. Editing with Geneative Edit results in a resized image up to 12MP. A visible watermark is placed on the image output upon saving to indicate that the image was generated by AI. The resolution of The reliability of the resulting output is not guaranteed.
Samsung did not immediately respond to Gizmodo’s request for comment if it was aware of these issues and if it was working on a fix. The Object Eraser tool, just like Google’s Magic Eraser before it, replaces the pixels of unwanted objects in the image to blend them into the background. It remains unclear why the system does not replace the same watermark in the new image.
Big AI companies love Adobe And Even the White House It promoted watermarking as a way to deal with the influx of AI deepfakes across the Internet. These watermarks, including Samsung’s, also provide metadata in each image, letting you know that the image has been modified. However, editing image metadata with a little knowledge is very easy. Photos uploaded to apps like Instagram will not display any metadata to users who unconsciously flip through their accounts.
then again, Artificial intelligence has long proven its ability to remove watermarks. If Samsung fixes the issue and makes watermarks more waterproof, it won’t stop users from taking the same photo in any other similar apps to remove them. Google’s DeepMind previously announced that it has a way to do this Install a more permanent watermark within the pixels themselves From a picture. DeepMind didn’t offer any insight into how this would work or what tools people could use to identify these altered images, but the company stated that the feature will only work on images modified with Google’s own tools.
For the rest of us, we’ll have to make do with half measures.
(Tags for translation)Samsung