Google Allows Users to Strip Visible AI Watermarks
Google is rolling out a new update that lets users remove visible watermarks from AI-generated images, videos, and songs. The change spans platforms including Gemini and the video editor Flow, while quietly retaining invisible SynthID watermarks and C2PA metadata.
Rollout Details for Nano Banana, Omni, and Lyria
The visible watermark toggle applies to specific generative models including Nano Banana, Omni, and Lyria. Josh Woodward, Google’s VP for Gemini, confirmed in a post on X that the setting is rolling out across Gemini and the Flow video editor. Search integration will arrive soon. Users can access the feature by navigating to Settings > Media Watermark once the rollout reaches their account.
Preserving Provenance Without Distracting Graphics
Industry standards for AI safety typically rely on visible branding to alert audiences. Yet, creators frequently argue that stamps obstruct professional and publishing workflows. Google maintains transparency by ensuring that invisible SynthID watermarks and C2PA metadata remain intact despite the removal of visible stamps. This dual approach allows platforms and consumers to check content origins via Gemini or Search without forcing creators to display distracting graphics.
Credentio Open-Source Library and Industry Compliance
Alongside the consumer feature rollout, Google is open-sourcing a library named Credentio. The tool helps developers integrate local validation mechanisms directly into their applications. This infrastructure update follows similar moves across the tech sector, such as Anthropic’s implementation of text and file watermarking in Claude to comply with European Union regulations, highlighting a broader industry push to standardize AI content labeling.
Worth a look