Google allows removing visible watermarks from AI-generated content as Anthropic takes different approach

Google announced that users can remove visible watermarks from images, videos, and music generated by its AI models (Nano Banana, Omni, Lyria), while maintaining invisible SynthID watermarks and C2PA metadata for traceability. The company also open-sourced a library called Credentio to enable local validation in applications, offering creative control while maintaining transparency. The move contrasts with Anthropic's recent decision to add watermarks to text and files generated by Claude to comply with European Union regulations, which has drawn controversy among corporate and educational users.