For years, the website contained information about what crawlers were not permitted on the site using the Robots.txt file. Adobe, who wants to create a similar standard for images, has added a tool to include credentials, with the intention of giving you a little more control over what is used to train AI models.
The main challenge may be to convince AI companies to actually adhere to Adobe standards. It’s already known, especially given that AI crawlers are already known to ignore requests for Robots.txt. file.
Content credentials are the metadata information for media files used to identify reliability and ownership. This is a type of coalition implementation for content origin and reliability, and is the standard for content reliability.

Adobe is releasing a new web tool that allows creators to attach content credentials to all image files, even if they are not created or edited via their own tools. Additionally, it provides a way for creators to signal AI companies that creators should not use that particular image in their training models.
Adobe’s new web app, called the Adobe Content Authincity App, allows users to attach credentials, including their names and social media accounts, to files. Users can attach these credentials to up to 50 JPG or PNG files at a time.
Adobe is partnering with LinkedIn to utilize a platform verification program owned by Microsoft. This helps to prove that the person attaching the credentials to the image has a confirmed name on LinkedIn.

Users can also attach Instagram or X profiles to images, but they are not integrated with validation of these platforms.
Using the same app, users should check the box and do not use images for model training.
This field is present in the app and then in the image metadata with content credentials, but Adobe has not agreed with any of the AI modelers to adopt this standard. The company said it is meeting with all top AI model developers to discuss this standard to persuade them to use and respect it.
Adobe’s intent is in a good place to provide indicators that model the makers of AI training data, but if companies don’t agree with the standards or respect the indicators, the initiative won’t work.

Last year, the implementation of labels on Meta’s platform caused a fuss as photographers complained that the edited images were tagged with “made with AI” labels. Meta later changed the label to “AI Info.”
This development highlighted that Meta and Adobe are both part of the C2PA steering committee, but there are differences in implementation on various platforms.
Andy Parson, senior director of Adobe’s Content Authenticity Initiative, said the company has built new content credentials with Creators. Given the global spread of regulations on copyright and AI training data, the company wants to provide creators with a way to use apps to demonstrate their intentions about the AI platform.
“Content creators want an easy way to show they don’t want to use their content for Gen AI training. We’ve heard from small creators and agencies that they want to have more control over their work. [in terms of AI training on their content]Person told TechCrunch.
Adobe has also released a Chrome extension that allows users to identify images with content credentials.
The company embeds the metadata in its content credentials app into various pixels of the image using a mix of digital fingerprints, open source watermarks and cryptographic metadata, so even if the image changes, the metadata remains intact. This means that users can use the Chrome extension to check content credentials on platforms such as Instagram that do not natively support the standard. Users will see a small “CR” symbol in the image when content credentials are attached.
In a world where there is a lot of debate about AI and art, Parson says he doesn’t believe in the C2PA to give views or direct what art is. However, he believes that content credentials can be a key marker of ownership.
“There’s a gray area [of when an image is edited using AI, but it is not 100% AI-generated]and what we’re saying is that artists and creators can sign their work and assert their belongings. This doesn’t mean that the IP is legal or copyrighted, it shows that someone made it,” said Parson.
Adobe also wants to add video and audio support while the new tools are designed for images.
Source link