Google is giving Gemini users more control over AI-generated media by introducing an option to hide visible watermarks on images, videos and music. The setting is being rolled out over the next few days and allows users to switch the “Show watermark” option on or off for future creations.
The feature is available under Gemini Settings as “Media Watermark”, with separate On and Off options. However, visible watermarks will continue to appear in countries where displaying them is legally required.
Josh Woodward, Google’s vice president leading Gemini, AI Studio and Google Labs, said the option will also come to Flow, followed by Google Search. He said the change was driven by user feedback and was among the most requested changes in a survey of Gemini complaints.
Invisible AI markers will remain
Disabling the visible watermark will not remove Gemini’s other provenance markers. Google said invisible SynthID watermarks and C2PA metadata will continue to be embedded in generated content. These markers can provide information about the AI origin of media even when the visible watermark is removed.
Woodward said the approach is designed to balance creative control with transparency. Users can also use Gemini or Search to check whether an image was generated using AI.
Google is also open-sourcing Credentio, a C++ library for working with C2PA Content Credentials. The library supports local validation of provenance information without requiring media files to be sent to cloud servers or external validation services.
According to Google, Credentio is based on code used across nearly 40 C2PA-enabled products and has supported billions of generated assets, including images, videos, audio files and documents. It currently focuses on validating C2PA credentials and can examine manifests, digital signatures and other embedded claims.
Anthropic adds invisible markers to AI text
The development comes as other AI companies adopt invisible markers for generated content. Anthropic recently said new Claude models will embed machine-readable watermarks in generated text that can survive copying and some editing.
Anthropic said the system works at the model level and serves as a provenance signal rather than a definitive AI detector. It applies to supported Claude models across Claude, the API, Claude Code, Claude Cowork and Claude Tag, as part of its commitments under the European Union’s AI Act Article 50(2) Code of Practice on Transparency of AI-Generated Content.
Also read: Viksit Workforce for a Viksit Bharat
Do Follow: The Mainstream LinkedIn | The Mainstream Facebook | The Mainstream Youtube | The Mainstream Twitter
About us:
The Mainstream is a premier platform delivering the latest updates and informed perspectives across the technology business and cyber landscape. Built on research-driven, thought leadership and original intellectual property, The Mainstream also curates summits & conferences that convene decision makers to explore how technology reshapes industries and leadership. With a growing presence in India and globally across the Middle East, Africa, ASEAN, the USA, the UK and Australia, The Mainstream carries a vision to bring the latest happenings and insights to 8.2 billion people and to place technology at the centre of conversation for leaders navigating the future.