Now OpenAi Made It Easy to tell you that an images is generated by DALL-3

 This is the first step toward making it easier to tell if a picture was taken with the DALL-E 3. This week, the company announced that it will soon start adding two types of watermarks to all pictures made by DALL-E 3. This is to meet the standards set by the C2PA (The Coalition for Content Provenance and Authenticity). The change has already been made to pictures made on the website and through the API. On February 12, mobile users will start to see the watermarks.

The first of the two watermarks is only in the image's information. You can use the Content Credentials Verify website and others like it to look at the information about how a picture was made. In the upper left part of the picture, you'll be able to see the second watermark, which is a CR symbol.

It's a good change that makes DALL-E 3 go in the right way and lets people know when something was made with AI. Similar watermarking is used in the metadata by other AI systems. Google has its own watermark to help identify pictures made with its image generation model, which was recently added to Google Bard.

As of this writing, the watermark will only be on still pictures. There will still be no watermarks on the videos or text. The company OpenAI says that adding the watermark to the metadata shouldn't cause any latency problems or lower the quality of the images that are made. It will, however, make images a little bigger in some jobs.

No matter how helpful watermarking is, it is not a foolproof way to stop false information from spreading through AI-made content. Because screenshots are used, metadata can still be left out, and most watermarks can be cut out of pictures. But OpenAI thinks that these methods will help users understand that these signs are important for making digital information more reliable and that they will be less likely to abuse the systems it has set up.