Google Adds Toggle to Remove Visible Watermarks from AI Content
Google now allows users to disable visible watermarks on AI-generated images, videos, and music in Gemini and its video generator Flow. A new 'Media watermark' setting removes the sparkle icon from the bottom-right corner. The update gives creators more control over the appearance of their AI outputs.
The new setting, located within Gemini's configuration menu, removes the sparkle icon from outputs generated by Nano Banana and Omni models. While this visual marker disappears, the underlying invisible SynthID and C2PA metadata remain intact, allowing users to query Gemini or Search to confirm an item's synthetic origin.
This adjustment aligns Google with competitors like OpenAI and Meta, which already avoid visible markers. Anthropic recently introduced invisible watermarking for text and images. Google will extend the toggle to Search, though it will refrain from doing so in jurisdictions where visible watermarks are legally mandated.
Removing visible markers could complicate casual detection of synthetic media, potentially increasing the burden on average viewers to verify content authenticity. However, since invisible metadata persists, platforms and researchers may still trace provenance. This shift could affect how creators present AI work, allowing for cleaner aesthetics, while simultaneously raising concerns about misinformation if users fail to utilize verification tools. Regulators in certain regions may resist this change, given existing transparency mandates.