Google Lets Users Remove AI Watermarks
Google has introduced a new setting that lets users turn off visible watermarks on AI-generated images, videos and songs. The change applies to content created through Google’s Gemini tools and selected generative AI models.
The move gives creators greater control over how generated media appears in final outputs. However, Google will continue embedding invisible identifiers, including SynthID and C2PA-related metadata, to preserve content provenance.
Visible Watermarks Become Optional
Previously, Google’s AI-generated media displayed a visible sparkle-style watermark that indicated its AI origin. Now, users can disable that mark through a new Media Watermark setting in supported Google products.
The option is particularly relevant for creators who use AI-generated material in professional projects. A visible watermark can affect the appearance of images, videos or other creative assets. Therefore, removing it can give users more flexibility when preparing content for publication or production.
The setting applies to AI-generated images, videos and songs. Google is rolling out the option through Gemini and its Flow video creation platform, while support for Search is expected to follow.
However, the visible watermark remains enabled by default. Users who want to remove it must change the setting before generating new content.
SynthID Remains Embedded
Despite the visible change, Google is keeping its underlying provenance technology intact. The company says invisible SynthID watermarks will continue to identify content generated or edited by its AI systems.
Moreover, C2PA metadata will remain available as another layer for documenting digital content’s origin and history. Consequently, removing the visible mark does not make the generated file completely free of AI-related identification.
Google’s SynthID technology embeds imperceptible signals directly into AI-generated images, audio, video and text. These signals are designed to remain detectable after common changes such as compression, resizing and colour adjustments.
Users can also check AI-generated media through Gemini’s verification tools. The system can scan supported images, videos and audio for SynthID and provide information about whether Google AI helped create or edit the content.
Balancing Creative Control and Transparency
Google’s decision reflects a broader debate over how AI-generated content should be labelled. Visible marks make AI origins easier to spot, while invisible identifiers provide a technical method for verification.
At the same time, creators have argued that visible watermarks can interfere with professional use. Google is therefore separating the visible label from the underlying provenance system.
Josh Woodward, Google’s vice president for Gemini, AI Studio and Google Labs, said the company is “striking a balance here between creative control and safety.” He added that invisible SynthID watermarks and C2PA metadata will continue supporting transparency.
The change could make Google’s AI tools more practical for creative workflows. Nevertheless, users and platforms will still have mechanisms to verify whether supported media originated from Google’s AI systems.
