Google to allow removal of visible watermarks on AI-generated images and videos

Jakarta (ANTARA) - Users of Google’s artificial intelligence (AI) services will be able to remove visible watermarks or watermarks from images and videos, Google Vice President Josh Woodward said in a social media post.

The option to add and remove watermarks will be available for the Nano Banana, Omni and Lyria AI models. In these three AI models, content is usually marked by Google with a familiar sparkle icon to indicate AI-generated content, according to a report by Endgadget on Friday (14/8).

The feature is also set to be available in Gemini and Google’s video editor service called Flow.

The markers in question are SynthID and C2PA metadata. Both may be difficult for ordinary users or the general public to identify, and they would need to dig deeper to check for them.

Woodward said the option to remove watermarks on Google-made AI content is being released in almost all countries except “in countries that are required by law to retain them”.

This means the feature will not be available in the European Union, which recently announced rules on AI transparency and requires watermarks to be visible on AI content that “appears authentic”. South Korea has also taken similar steps in early 2026.

Watermarks sometimes become an obstacle for users who use AI to assist with creative work.

“We are balancing creative control and safety,” Woodward said.

Markers for AI content are currently a focus of attention for many parties globally. Besides Google, Suno recently offered an AI service for making music and announced a labelling system to show that its content is indeed made by AI.

The labelling can also serve to show that AI-generated content can be identified by other audio platforms as well.

In addition, Anthropic also announced that it has added invisible C2PA-based AI content markers to text content or files generated by Claude.