Dall-E 3 will add authenticity tags to all AI-generated images

Feb 12, 2024

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Dall-E 3 will add authenticity tags to all AI-generated images

Feb 12, 2024

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Join the Discussion

Share on:

Dall-E 3 will add authenticity tags to all AI-generated images

OpenAI is adding watermarks to all images created using Dall-E 3. The step is to encode a content authenticity tag in all AI-generated images, which viewers can look up to check the origins of the image.

OpenAI, who is behind both Dall-E and the popular ChatGPT, says that this move is to help build public trust in digital information. And in a big US election year, that is not a moment too soon.

OpenAI is implementing the watermarks in the metadata as part of a universal push to standardise the Coalition for Content Provenance and Authenticity (C2PA). Watermarks from C2PA will appear in images generated on the ChatGPT website and the API for the DALL-E 3 model.

As of today, the watermarks will also appear on images generated by mobile users as well as the desktop version. The information will be stored in the metadata, and a visible CR symbol will appear in the top left corner of each image. At the moment, the watermark is only available on still images, not videos.

Dall-E 3 will add authenticity tags to all AI-generated images

To check the authenticity of an image, viewers can go to an independent website such as Content Credentials Verify to check where it was made.

Camera brands

So far, companies such as Adobe and Microsoft have adopted the C2PA tag to help differentiate between AI-generated content and human-made art. This is a positive step in reducing deepfake images and videos in an important political year for many countries.

Meta is also expected to join by tagging any content made with their AI features. Several camera brands, such as Nikon, Sony and Leica, have also joined the scheme.

Deepfakes

After Taylor Swift was the unfortunate latest victim of explicit deepfake content, several countries are moving to impose legislation making it a criminal act. Even the Pope made a forceful argument for tackling the issue in his latest public address.

Although this is certainly a step in the right direction, this isn’t one hundred per cent foolproof. Metadata can still be tampered with and changed or deleted. Additionally, taking a screenshot of an image voids any metadata that could be contained, and of course, we all know how easy it is to remove watermarks.

[via the verge]

Find this interesting? Share it with your friends!

Alex Baker

Alex Baker

Alex Baker is a portrait and lifestyle driven photographer based in Valencia, Spain. She works on a range of projects from commercial to fine art and has had work featured in publications such as The Daily Mail, Conde Nast Traveller and El Mundo, and has exhibited work across Europe

Join the Discussion

DIYP Comment Policy
Be nice, be on-topic, no personal information or flames.

Leave a Reply

Your email address will not be published. Required fields are marked *