Adobe is stepping up its efforts to protect creators’ digital work with a new web-based tool designed for securing attribution and safeguarding content against unauthorized AI usage. Launch, the Adobe Content Authenticity app will allow users to attach detailed metadata, known as Content Credentials, to digital files like images, audio, and video.
These credentials act as a “nutrition label” for media, containing details such as the creator’s name, website, and other identifying information, helping to verify ownership and maintain content integrity.
Addressing AI-Related Concerns
Creators have increasingly voiced concerns over the use of their work in training AI models without permission. According to Adobe, more than 90% of creators are looking for a standardized method to attach attribution to their content, and over half express apprehension about their work being used in AI development without their consent.
Adobe’s new app aims to meet this demand by providing a streamlined way to apply Content Credentials in bulk, offering better protection for creators who want to keep control over their intellectual property. The web app, expected to enter public beta in Q1 2025, will be available for free and won’t require an Adobe subscription.
It integrates with Adobe’s existing suite of Content Credentials tools, which are already supported in products like Photoshop, Lightroom, and Firefly, and serves as a central platform where creators can manage these preferences. The ability to attach credentials isn’t limited to Adobe software; creators can add metadata to any content, regardless of the software used to produce it.
Staying Connected Even After Edits
The new Content Credentials system is more resilient than traditional metadata. Adobe incorporates a mix of invisible watermarking, cryptographic signing, and digital fingerprinting, making it possible to restore attribution details even if a file’s metadata is stripped away or when a screenshot is taken. The added durability is aimed at helping creators maintain control over their work’s provenance throughout its lifecycle.
Adobe is also using the app to promote a preference setting that enables creators to indicate whether their content can be used in training generative AI models. While Adobe’s Firefly AI models only train on licensed content, not all generative AI systems follow similar guidelines.
By opting out through the app, creators can signal to AI developers that their work should not be included in datasets used to train models. Currently, Spawning, a company known for its AI training database tool “Have I Been Trained?” has committed to recognizing these opt-out preferences, but broader industry support remains limited.
For creators concerned about unauthorized AI usage, the app provides a central hub to manage their preferences, potentially setting the stage for more consistent practices in the AI and tech sectors. Adobe continues to advocate for wider acceptance of these practices, hoping that major players such as OpenAI and Google will adopt similar opt-out options.
Bringing Transparency to Online Content
One ongoing issue is that many social media platforms and websites do not display metadata that includes Content Credentials, limiting visibility for viewers. To help bridge this gap, Adobe is introducing a Chrome extension that allows users to inspect any Content Credentials embedded in media on the web.
The app also features an inspection tool that can display the editing history and other information associated with a piece of content when available. These tools aim to make provenance details accessible to a wider audience, offering an additional layer of transparency.
As the company prepares to debut the app at Adobe MAX, the world’s largest creative conference, from October 14-16 in Miami Beach, it will offer attendees a preview of how the new system can reshape digital content management. The beta version of the Content Authenticity app is expected to draw interest from a wide range of digital creators.
Last Updated on November 7, 2024 2:37 pm CET
