Visual storytelling—whether through art, design, or photojournalism—can help people change the world by sharing unique viewpoints, celebrating diverse communities, even spurring important social activism. Digital tools have empowered unprecedented creative potential in visual storytelling, but in the wrong hands they can also enable convincing deception. As artificial intelligence gives computers the ability to conjure up convincing synthetic imagery, telling stories authentically matters more than ever. The clearest path to a more trustworthy internet is digital provenance—a technology that provides simple, objective facts about how something was made, so we can experience context and content together.
At its core, provenance is a chain of trust signals, not a simple judgment of whether something is true. Provenance provides us with a layer of transparency that informs our understanding of how an image, video, or audio file came to be. Because each link in the chain is secure and bound to the content itself, provenance cannot be altered without leaving evidence of tampering. We can rely on its guarantee of transparency to make informed decisions about the content we see online.
This is the vision behind our Adobe-led Content Authenticity Initiative (CAI) and its supporting standards development organization, the Coalition for Content Provenance and Authenticity (C2PA). These efforts are advancing an industry-wide open provenance stack that embeds provenance throughout the digital content lifecycle. This ensures proper attribution for content creators as well as clarity about the origin and history of the visual stories we experience.
Imagine you encounter an image online and want to know more about it. You need only click an icon to see who made the image, where it came from, edits that have been made, other works that may have been incorporated, and whether it has been altered since the last link in the chain was added. Additionally, potentially sensitive signals like location and identity can be excluded or redacted later to protect privacy and security, while maintaining other provenance signals. A truly transparent content ecosystem is within reach once a provenance stack is deployed throughout the digital content lifecycle.
The provenance stack has three layers: open standards for attaching trust signals to media, open-source tools for easy integration into products, and user experiences that make provenance commonplace throughout the web, in mobile messaging, and even immersive media. Each part of the stack builds upon the last, forming a clear path to a more trustworthy internet. The provenance stack is real, available, and sees a major step forward today with its incorporation into cameras.
A milestone moment for authenticity in photo capture
In photojournalism, putting the provenance stack to work throughout the content lifecycle means starting from the moment an image is captured. CAI partnerships with the camera manufacturers Leica and Nikon, announced at Adobe MAX 2022, have made that critical first step possible. Photojournalists and prosumer photographers alike will soon be able to securely attach information like their name, date, location, and time of image capture to their photos. Starting in 2023, when Leica plans to make provenance features available, we’ll see the first broad use of the provenance stack by visual storytellers.
Additionally, Nikon is developing C2PA functionality for a future camera model. As the maker of cameras used by journalists at leading news agencies, including fellow CAI member Agence France-Presse (AFP), this represents a significant opportunity to enhance the power of journalistic storytelling by providing authenticity at the source. These cameras and the provenance feature in Photoshop called Content Credentials comprise an important step for the provenance stack: rapidly growing availability to all.
Think of the wartime photographer on the ground, who can now show that their eye-opening images of current events were, in fact, taken in the conflict zone. Or the art photographer who can add an indelible record of attribution. These storytellers benefit from having verifiable context attached to their images, and their audiences benefit from easy access to it. The applications of provenance in photography are many and, as with any foundational technology, new ones will emerge.
Perhaps the most important aspect of the provenance stack is that it empowers the public in the face of a rising tide of misinformation. If we have the means to verify how a piece of digital content came to be, we can make informed decisions about whether to trust what we see. And when content lacks provenance, we may justifiably raise questions about it.
From vision to reality: CAI progress
Since our founding, the CAI has made steady progress. In just three years, we have grown our membership to 800, including some of the world’s leading newsrooms, software companies, hardware companies, and NGOs, all united behind the goal of restoring authenticity and bringing transparency to the digital ecosystem through the provenance stack.
The user experience part of the provenance stack already exists in Adobe tools like Photoshop, Adobe Stock, and Behance. As the next step in the digital content lifecycle after image capture, editing software like Photoshop is a place where provenance must be incorporated. Imagine our wartime photographer from earlier. Now, when they send their photo to an editor back at the news desk, the photo’s origin information travels with it. When the editor brightens the image or makes some minor cropping decisions, those edits are added to the photo’s history – building upon the secure chain of trust signals the CAI seeks to enable from capture all the way to sharing on social media.
The C2PA standards organization that provides the base of the provenance stack is led by Adobe, Arm, BBC, Intel, Microsoft, Sony, Truepic, and Twitter. In January, the C2PA released its open technical standard for provenance, and refinements are being added quickly. The CAI open-source suite of tools was released in June and provides the middle layer in the provenance stack. It is free, open, simple-to-use code enabling developers to build websites, services, mobile apps, and even embedded systems, as the Nikon team has done.
Championing a future backed by provenance
With these open-source tools and standards, we will see provenance technology incorporated into other components of the digital content lifecycle. And as storytelling continues to evolve, provenance will bring authenticity to new mediums, styles, and techniques, including generative AI. Generative AI is a revolutionary class of computing that can help unleash imagination. As the marriage of AI and human creativity unfolds, there will be value for artists to show how their work was created. The provenance stack can provide attribution so creators can continue to tell their stories in a transparent and authentic way.
Authenticity matters. It matters to artists, photojournalists, and responsible developers of powerful creative technologies. Over the past three years, the CAI has put in place a foundation for trustworthy storytelling from the instant of capture to the moment of visual experience. It is now time to harness the potential of provenance to ensure every storyteller can enrich culture and shine a light on the world, authentically.
This story was written by Andy Parsons, Sr. Director of the Content Authenticity Initiative at Adobe and produced by WIRED Brand Lab.
