Google now lets you create AI images without the annoying watermark in the corner
Google is introducing a 'Media watermark' setting that allows users to disable the visible AI-generated icon. Despite this, hidden identification methods like SynthID and the C2PA standard will remain embedded in all generated files.
One of the ways to quickly know if an image is real or generated by AI is the "watermark" that Google added to the bottom right of the image. However, that sparkle icon also annoyed users, who found it easy to remove with Photoshop or a quick crop, but wasted time doing so. Now, Google is changing direction and allowing users to remove this mark in advance.
It wasn't really effective
Google announced that in the coming days, a new setting called Media watermark will be added to the Gemini app as well as its video generator, Flow. Once you turn off the switch in the settings menu, the sparkle watermark will stop appearing, making it difficult to identify the image as AI-generated at a glance. The new setting will apply to the company's models, such as Nano Banana for images, Omni for video, and Lyria for music clips.
Of course, the absence of a corner watermark does not mean the image cannot be identified as AI-generated. In fact, content identification has long relied on information embedded within the image file rather than visible signs.
Google emphasized that hidden identification mechanisms will continue to operate as usual:
"The SynthID technology, which embeds an invisible digital stamp, along with C2PA, an open standard that embeds metadata in the file itself, will remain an integral part of generated files."
If you want to verify if an image was created by AI, you can simply drag it into Gemini and ask for a check. Josh Woodward, Vice President for Gemini at Google, explained the move on X, noting that the company is trying to find a balance between "creative control" and safety. Visible watermarks are becoming optional, but hidden technologies will continue to maintain transparency.
In parallel, Google is releasing an open-source library called Credentio, which will allow developers to integrate local verification mechanisms directly into their applications or websites.
Google aligns with competitors
With this step, Google is joining competitors like OpenAI and Meta, which did not initially embed visible watermarks in their generated content. Like Google, OpenAI has relied for years on SynthID and the C2PA standard, while Meta uses its own. However, earlier this week, Anthropic announced it would start embedding "watermarks" in texts and files generated by Claude to meet strict European Union regulatory requirements.
The option to remove the watermark is currently reaching users. However, in countries that require clear and visible AI-generated content labels, this setting will not appear.
