Google allows users to remove visible watermarks from AI-generated content

Google is rolling out a new setting on Gemini and its Flow video editor that allows users to remove visible "sparkle" watermarks from AI-generated images, videos, and music. While this change grants creators more flexibility, Google maintains safety standards by keeping invisible SynthID watermarks and C2PA metadata embedded in the files. The toggle is restricted for school and work accounts, requires a Google AI Ultra subscription in India, South Korea, and Vietnam, and is unavailable where visible watermarks are legally required.

Visible watermark removal option

• ▪Google announced on August 14, 2026, that it is rolling out a new setting allowing users to toggle off visible watermarks on images, videos, and music generated through Gemini Apps and the Flow video editor.
• ▪The visible watermark toggle applies to content generated using Google's Nano Banana, Omni, and Lyria AI models, with support for Google Search coming soon.
• ▪When the visible watermark is toggled off, Google removes the "sparkle" watermark that typically appears in the bottom-right corner of generated content.

Invisible SynthID watermarks

• ▪Users can still verify if an image was AI-generated by using Google Gemini or Google Search to detect the invisible SynthID watermarks.
• ▪Turning off the visible watermark does not remove Google's invisible SynthID watermark, which embeds an imperceptible signal into generated media to identify if it was created or edited using Google's AI.

C2PA metadata embedding

• ▪AI-generated media created or edited through Gemini will continue to contain Content Credentials based on the C2PA standard even if visible watermarks are disabled.
• ▪Google is open sourcing a new C++ library called Credentio to allow developers to embed a local C2PA validation mechanism in their own applications.

Regional subscription restrictions

• ▪Users in India, South Korea, and Vietnam can only access the visible watermark toggle if they have a Google AI Ultra subscription.
• ▪The visible watermark toggle will not launch in countries that legally require visible watermarks, and non-subscription users in India, South Korea, and Vietnam will continue to have visible watermarks automatically applied.

Work account limitations

• ▪The visible watermark toggle is unavailable for users accessing Gemini through work or school accounts, meaning visible watermarks will remain enabled for those accounts.
• ▪The new watermark setting applies specifically to Gemini Apps on web and mobile and does not change watermark policies for AI features offered through other Google products.

Industry watermarking practices

• ▪Other AI image generators, such as OpenAI and Meta, do not apply visible watermarks, relying instead on standards like SynthID, C2PA, and Meta's Content Seal.
• ▪Anthropic announced in August 2026 that it is applying watermarks to both AI-generated text and images to comply with European Union regulations.