Google Introduces Feature to Eliminate Visible Watermarks from AI-Generated Content
Invisible watermarks, however, will still be intact.
In a recent announcement from Google, Vice President Josh Woodward revealed that users will have the option to remove visible watermarks from images and videos created by AI. This feature will soon be available for the Nano Banana, Omni, and Lyria models, recognizable by their typical sparkle icon indicating AI usage. Future updates will also extend this functionality to Gemini and the Flow video editor.
The decision raises concerns about the effectiveness of visible watermarks, designed to help people easily identify AI-generated content. Despite this, there will still be alternative methods to detect such content. Crucially, the invisible SynthID watermark and C2PA metadata will remain unaffected by this change, as there are no options to disable them. However, accessing these will require users to engage with a chatbot or submit content through a specific portal.
Woodward noted that the watermark toggle feature is set to launch widely, “except in regions where regulations mandate their retention.” This includes the European Union, which has introduced new regulations focusing on AI transparency that require visible watermarks on realistically produced deepfake content. South Korea has also implemented similar legislation earlier this year.
Google’s decision reflects the rapidly changing landscape of how AI-generated content is labeled. While visible watermarks serve as an effective tool for helping average users recognize deepfakes, they can also pose challenges for creators utilizing AI in their work. As Woodward stated, “We are aiming for a balance between creative freedom and user safety.”
This update is part of a broader trend concerning watermarks in the AI domain. Recently, Suno introduced new “transparency tools,” featuring a labeling system that indicates whether a song has been produced by AI when shared on other platforms. Additionally, Anthropic has announced plans to implement invisible C2PA-based watermarks on text and files produced by Claude to ensure compliance with the aforementioned EU regulations.
