OpenAI decided to add watermarks to the images created on the ChatGPT website and API for the DALL-E 3 model, reports The Verge.

This means that the company will join the tech giants that are deploying support for the Content Provenance and Authentication Coalition (C2PA) standards. This currently includes companies such as Adobe and Microsoft.

Mobile users are expected to see watermarks on AI-generated images by February 12. It is a visible CR symbol in the upper left corner of the image. But the watermark will also have an invisible metadata component.


People will be able to verify the origin of content on websites such as Content Credentials Verify. At this stage, watermarks will only appear on still images. So video or text is out of the question for now.

OpenAI warns that such marks will have a minor impact on the time it takes to create images, but not on their quality. The company also says that C2PA metadata can be “easily be removed either accidentally or intentionally.”

By the way, it has recently become known that Google is launching a new tool for creating images based on artificial intelligence. It is called ImageFX and is based on the Imagen 2 model.

The images created with it will be watermarked within the SynthID technology. This will serve as an additional security measure.