Adobe is moving deeper into the fight over AI, ownership, and trust online with a Content Authenticity web app designed to help creators mark their work as theirs. The beta is planned for the first quarter of 2025, and the company is also releasing a Chrome extension in beta on Tuesday.
The idea is straightforward: make creator credit and provenance harder to remove as images, videos, and audio files move across the web. The execution is more technical than ordinary file metadata, because ordinary metadata can disappear easily when content is copied, reposted, or captured through a screenshot.
How Adobe wants content credentials to work
The web app will let creators apply content credentials to their work. Those credentials are meant to certify ownership and help viewers understand where a piece of media came from.
Adobe’s approach uses three main methods together: digital fingerprinting, invisible watermarking, and cryptographically signed metadata. Each part addresses a different weakness in how digital media travels online.
- Invisible watermarking makes tiny pixel-level changes that people cannot detect by looking at the image.
- Digital fingerprinting encodes an ID into the file, so the work can still be identified even if content credentials are removed.
- Cryptographically signed metadata adds a stronger provenance layer than ordinary metadata alone.
That combination is meant to keep the connection between the creator and the file intact across images, videos, and audio files. Andy Parsons, Adobe’s senior director of Content Authenticity, told TechCrunch that the goal is for the content credential to remain attached wherever an image, video, or audio file goes on the web or on a mobile device.
Why adoption is the real test
A system like this depends on people and platforms actually using it. Adobe has an obvious advantage because it sits at the center of many creative workflows and has 33 million subscribers who pay for its software.
The company is also opening the tool beyond its existing customer base. Artists who do not use Adobe software will still be able to use the web app to add content credentials to their work.
That broader access matters because AI-related content disputes do not only affect professional users inside Adobe’s ecosystem. Images and other media can be copied, reposted, edited, and mixed into new contexts long after they leave the original creator’s hands.
Adobe has also co-founded two industry groups focused on content authenticity, trust, and transparency online. Their membership includes camera manufacturers representing 90% of the market, content-creation tools from Microsoft and OpenAI, and platforms including TikTok, LinkedIn, Google, Instagram, and Facebook.
Membership in those groups does not guarantee that each company will build Adobe’s content credentials into its own products. But it does give Adobe a route to push the idea across companies that shape how content is created, distributed, and viewed.
Making provenance visible across the web
Even if credentials are attached to a file, people still need a way to see them. The source article notes that not every social media platform or website visibly displays provenance information.
Adobe’s answer is to provide discovery tools. The Content Authenticity Chrome extension is part of the software package, and Adobe is also offering an Inspect tool inside the Adobe Content Authenticity website.
Those tools are intended to find and display content credentials wherever they are associated with content online. In practical terms, that could help viewers see who made a piece of content and who should receive credit for it.
This is especially important because AI itself is not reliable at determining whether something is AI-generated. As synthetic images become harder to distinguish from real ones, credentials offer a different path: identifying origin and attribution when the media has been marked in the first place.
Adobe’s position on AI and creator consent
Adobe is not presenting this as an anti-AI move. The company has its own generative AI tool, Firefly, and the source article says Firefly is trained on Adobe Stock images.
Parsons said Firefly is commercially safe and that Adobe trains it only on content it explicitly has permission to use, never on customer content. That distinction is central to Adobe’s message: the company wants AI use to be visible and creator permission to matter.
Adobe is also trying to prevent artists’ work from being used in training datasets without consent. That concern sits at the center of the wider backlash from artists who have resisted AI tools.
At the same time, Parsons said Firefly integrations in apps such as Photoshop and Lightroom have received positive feedback. He also said Photoshop’s generative fill feature, which can extend images through prompting, saw a 10x adoption rate over a typical Photoshop feature.
Where Spawning fits in
Adobe has also been working with Spawning, a tool focused on helping artists control how their work is used online. Spawning runs a website called Have I Been Trained?, where artists can search to see whether their artworks appear in the most popular training datasets.
Artists can also add their works to a Do Not Train registry. That registry signals to AI companies that the work should not be included in training datasets.
Like content credentials, the registry depends on cooperation. The source article notes that it is only effective if AI companies honor the list, and says Hugging Face and Stability are on board so far.
The common thread is clear: Adobe is trying to build practical infrastructure around attribution, provenance, and consent. Its tools cannot solve every problem created by AI-generated media and online copying, but they give creators another way to assert ownership and give viewers another way to check where content came from.