Google said Friday it will let users switch off visible watermarks on AI-generated images, video and songs, while retaining invisible SynthID watermarks and C2PA metadata for provenance. The option will roll out in coming days across Gemini and Google’s Flow video editor, with Search support to follow, and will apply to the company’s Nano Banana, Omni and Lyria models. Users can toggle the feature via Settings > Media Watermark. Google is also open-sourcing a library called Credentio to enable local validation in third-party apps. The move highlights ongoing industry trade-offs between creative utility and content transparency, and follows Anthropic’s decision to add watermarking to Claude outputs amid European regulatory pressure.
Related articles:
Coalition for Content Provenance and Authenticity (C2PA) — standards for content labeling
Adobe Content Credentials (CAI) — provenance for digital media
Digital watermarking
Artificial Intelligence Act (European Union)
