Samsung Specifies AI Watermark On Galaxy S24 Series

AI Watermark On Galaxy S24 Series

Samsung Specifies AI Watermark On Galaxy S24 Series, Avoiding Deepfake Controversy [DD Electronics]

Samsung Electronics announced that it will display an AI watermark on its new Galaxy S24 series. This is to avoid the recent deepfake controversy and respond to AI regulation laws. The Galaxy S24 series, which was unveiled on the 18th, is equipped with generative AI, including Galaxy AI, an on-device AI, in the smartphone. Functions where personal security is important, such as interpretation calls, are operated by on-device AI, and tasks that require advanced computation are handled by generative AI.

Photo editing using ‘creative editing’ installed in the Galaxy S24 series. (From left) When I change the positions of the person and the ball in the photo at the bottom, a loading time is displayed along with the message ‘Drawing a new landscape.’ The third photo is a composite image of a natural background screen created by generative AI.

Among the various AI functions, the ‘creative editing’ function had a notable caveat. ‘All images created based on generative AI are watermarked both in the image itself and in metadata (attribute information).’ Generative editing is a function that allows smartphone users to freely edit photos. You can move the position of the subject in the photo and adjust its size, but the photo may become crooked or have empty spaces during the editing process.

At this time, generative AI, which has learned numerous images, predicts that there will be an object in the empty space and fills it with images that match the surrounding empty space. A Samsung Electronics official said, “Generative AI creates images that did not exist before,” adding, “Recently, deepfakes, which synthesize images based on AI, have caused several controversies, and have the potential to be abused, so the Galaxy S24 ’s generative AI “A watermark is placed on images created with.

” He continued, “There is a regulation in China that requires watermarks to be provided to AI creations, and I understand that Europe will also implement AI- related regulations in about two years.” He added, “There are no related regulations yet in Korea, but each country has created generative AI.

“As discussions about images continue, Samsung has prepared a clear AI watermark,” he added. AI watermarks are always displayed on images created by generative AI, but the watermark can be erased at the user’s discretion. Foreign media also gave positive reviews to Samsung Electronics for specifying the AI ​​watermark and preemptively responding to the deepfake controversy and AI legislation.

Ben Wood, senior expert at CCS Insight, an IT analysis company, told CNBC on the 17th (local time), “Deepfake content such as image and video manipulation created based on AI raises ethical issues,” and added,“ Images and metadata altered by AI “Adding a watermark to all of our data is a constructive move by Samsung, and we are confident that other companies will follow.”

See More:

Apple Begins Pre-Orders For Vision Pro in US – Product Spec Also Revealed

Davos Forum2024: Dystopian AI Vision Disappears at Davos

ChatGPT Used in Education: First Introduced at the University of Arizona

Galaxy S24 Beats iPhone: Roh Tae-moon Agrees With The Comment