Google has made a significant update to its policies regarding AI-generated content by allowing users to remove the visible watermark from their generated images, videos, and songs. This announcement was made public on Friday, with the company confirming that the invisible SynthID watermark and metadata linked to the C2PA standard will remain intact.
Josh Woodward, Google’s Vice President of the Gemini team, elaborated in a post on X about the new feature. It will be accessible for models such as Nano Banana, Omni, and Lyria. Users will be able to disable the visible watermark via settings in the Gemini platform and the company’s Flow video editor, with future support planned for Google Search.
This adjustment signifies a shift in how AI-generated media is labeled. While visible watermarks can hinder the utility of professional and creative outputs, there is still a pressing need for accountability in identifying AI-generated content. Woodward emphasized this balance, stating, “We’re striking a balance here between creative control and safety: while the visible watermarks are now optional, invisible SynthID watermarks and C2PA metadata are still being used for transparency. So you can still use Gemini or Search to see if an image was AI-generated.”
The rollout of this feature is expected in the coming days. Users can find the option to toggle the visible watermark by navigating to Settings > Media Watermark.
Moreover, Google announced it will be open-sourcing a new library named Credentio, allowing developers to incorporate a local validation mechanism for their applications.
This update follows a contentious move by Anthropic, which recently introduced a watermark in text produced by its AI models to align with EU regulations.
Editor’s Take
This development signifies Google’s attempt to balance creative freedom and the need for transparency in AI-generated content. Removing visible watermarks can enhance the usability of AI creations for various professionals while maintaining hidden identifiers ensures accountability. This could pave the way for more engaging content creation while adhering to necessary regulations.
Source: techcrunch.com
