Adobe has a recent tool to protect artists’ work from artificial intelligence

Adobe is expanding its content credentials “nutrition labels” to make it even easier for creators to attribute their work, identify what is and isn’t AI on the web, and protect their content along the way. It’s launching a free web app that will allow users to quickly apply creator information to images, videos and audio, and even opt out of generative AI models – at least for the AI ​​creators that support it.

Content Authenticity web app can be used to broadly apply attribution data to content that includes the creator’s name, website, social media pages, and more. It also provides creators with an easier way to opt out of AI training en masse, compared to the painstaking task of providing individual security for their content to each AI vendor.

The web app will act as a centralized hub for Adobe’s existing content credentials platform. Content authentication data is tamper-evident metadata that can be embedded in digital content to reveal who owns it, who created it, and whether artificial intelligence tools were used to create it. The web app will integrate with Adobe’s Firefly AI models, as well as Photoshop, Lightroom and other Inventive Cloud apps that already support individual content credentials. Importantly, the hub will enable creators to apply content credentials everyone image, video and audio file – not just those created using Adobe applications.

The Content Authenticity web app also allows users to set Generative AI preferences to protect their work from being exploited or trained by Generative AI models. Adobe’s first-party models are only trained on licensed or public domain content, but these protections are designed to be broadly applicable to models produced by other companies – as long as those companies support them. Right now the list is compact: just Spawninga startup best known for creating “Have I been trained?” AI training database checker, has committed to supporting this feature now. Adobe says it is “actively working to adopt this preference across the industry.” We can hope that AI vendors like OpenAI and Google, who already support credentials, will follow suit.

According to Adobe, these AI attribution and preference tags will be more arduous to remove because anything associated with content credentials can be restored using a combination of a digital fingerprint, concealed watermark and cryptographic metadata, even if someone takes a screenshot of the protected work, Adobe says. This isn’t a completely bulletproof solution, but it does mean that anyone who wants to bypass these protections will have to put in more work.

Adobe also makes it easier to check whether authentication credentials, such as “AI Info” Meta tags, have been used on websites that do not present this information to users. The Content Authenticity web app includes a validator that recovers and displays content credentials and edit history when available, and the Content Authenticity extension for Google Chrome, launching in beta today, will be able to check content directly on a website.