AI-generated media (photos, videos, and even audio) have vastly improved over the years. You've probably seen the "Will Smith eating spaghetti" meme/prompt. In 2023, the results looked like this. And now it looks like this:
The interesting thing is that this is the worst this technology will ever be. It will likely continue to improve from here, with tools like Google Veo 2 and OpenAI Sora already competing in AI-generated video. This progress has raised ethical concerns, especially regarding the role of AI-generated media in spreading misinformation.
In response, AI labs have started adding watermarks to AI-generated images. xAI does this with Grok by placing the logo at its images' bottom right corner. Bigger companies like Google, however, are taking a different approach with SynthID.
SynthID was introduced in August 2023 as a tool for watermarking and identifying AI-generated images. Its capabilities expanded in May last year to cover AI-generated text and video.
For images and videos, the technology works by embedding a digital watermark into the pixels of AI-generated images and each frame of AI-generated videos in a way that does not compromise visual quality.
Now, in a recent announcement, Google says this watermarking technique is coming to Reimagine in Magic Editor. The company states:
Starting this week, Google Photos will begin using SynthID to mark images edited with generative AI using Reimagine in Magic Editor.
SynthID may not work for all edits made with Reimagine. For example, if you change the color of a small object in the background, the technology may not detect it and will not label the image as AI-generated.
Google says you can check if an image is AI-generated by using "About this image," a feature announced in 2023. It displays an appropriate label below AI-generated images. You can access it in Circle to Search by tapping or circling the image.
0 Comments - Add comment