OpenAI now offers watermarking technology for AI images created by DALL-E 3: How it works

OpenAI realizes that people need to be notified when they see an AI-generated image, especially one created by its own platform DALL-E 3, and the company provides watermarks for this purpose. OpenAI has confirmed that it will use open standards developed by the Content Provenance and Authenticity Alliance (C2PA). This means that any AI images generated using the DALL-E 3 will have metadata details, such as the name of the AI ​​tool used to create the image.

In an interesting coincidence, while Meta talks about policing AI-generated content and making its identity clear to viewers, OpenAI brings its own version of a watermark.

Since OpenAI is also integrating DALL-E 3 into ChatGPT to generate AI images, these new metadata details will be available to the AI ​​chatbot by February 12th. Depending on the details about the source of the image OpenAI may increase the file size, but this will not affect the visual quality, which may be more important to people.

It’s great to see OpenAI bringing these changes to AI-generated images, however, the company highlighted a major vulnerability in the feature that people can easily bypass.

OpenAI points to a trend of social media platforms removing metadata. You may also lose detail if you take a screenshot of said image, which can easily remove the source of the image, especially if they are AI-generated.

Situations like these demonstrate that strict regulation of AI content is critical and that all tech companies need to work together to make it safer to create and consume AI-generated images, videos, and other content.