A conversation with Santiago Lyon: Adobe Content Authenticity Initiative

Adobe last month launched a new Photoshop tool aimed at combating misinformation and bolstering image provenance. Part of the ongoing work being undertaken by the Adobe-led Content Authenticity Initiative (CAI) and its media partners, it assigns a cryptographic signature to a photo’s metadata and edit history. These “content credentials” ensure viewers can trust the image they see, while creators can detect stolen work. 

More broadly, the CAI is working to instate image provenance and attribution standards across the web. Major media players like the New York Times, Twitter (both founding members), Getty, AFP and the Washington Post have joined the CAI, as well as Microsoft, Qualcomm, Arm and a host of smaller organizations and individuals. 

We spoke with Santiago Lyon, Head of Advocacy and Education for the initiative, earlier this year.  He told us about the inspiration for the CAI, what organizations and individual visual storytellers stand to gain by joining, and how the software works.

TTW: So, how does that CAI tool work? 

SL: Adobe was challenged to react to this current climate of mis- and disinformation. And so, in its role as a good digital citizen, the company decided to put some resources behind coming up with solutions to the problem. 

So, we decided to develop and implement a very secure and robust form of metadata – what we’re calling content credentials.

When an image is created and then edited in Photoshop, or Lightroom, or Photo Mechanic, or any other program out there, this technology will allow any of those changes to be captured. And ultimately, when images are published, in both social and traditional media, we want to see them published with a  CAI logo or symbol that when moused-over or clicked on will expose some more information about the image’s provenance to the viewer.

TTW: Is the interest in the CAI coming primarily from organizations or from individuals?

SL: It’s coming from both. There’s a lot of interest in the topic, given the state of the industry and the threats that are out there and the increased prevalence of bad actors trying to influence things by falsifying images or modifying the context of images or whatever their methodology is. There’s a lot of interest at both the institutional and the individual level from a mix of media companies, photo editors, technology companies, academics and NGOs, in addition, within the creative community, artists, digital artists – who in many cases use Adobe software to create beautiful digital works of art – they’re interested too, as a way of tracking how their content is being used, (or abused in some cases), or allowing people to see how digital artists put their work together.

On an organizational level, I think the time will come when assigning organizations, will come to expect this from the photographers that they’re working with whether they’re working with staff photographers or freelancers, and it will be another tool in the photographer’s arsenal of credibility and veracity.

TTW: How does this look on social media? Has it been implemented yet? 

SL: A lot of the nefarious behaviour around content and authenticity happens on social media platforms. And so the key concept here is the integrity of the metadata. 

At the moment, when files enter content management systems or social media platforms, for the most part, the metadata is stripped off entirely. The viewer cannot see where the image came from. Was it pirated? Was it stolen? Was it a screengrab or original? Who’s the creator? Who’s the distributor? And so we’re working with Twitter and other social media players to explore the possibility of leaving this metadata intact, so that when a picture appears on a platform, there’s a way for the viewer to look into the provenance of the picture and see if it’s, in fact, linked to who the distributor is or where it came from originally, or what was done to it along the way.

We’re in the early stages with social media platforms. But there is keen interest on their part to participate in this endeavor.

TTW: What do you see in terms of how this progresses? What does five years from now look like? Ten years?

SL: When you look at each of the three areas that I mentioned – capture, edit, publish, we’re at different stages in each of those three areas but making steady progress.

It’s now in a beta version of Photoshop, and then we’ll turn our attention to Lightroom. And we’re also in conversations with Photo Mechanic and some of the other players in that space. 

But I think it’s a multi-year project, minimum, and that’s just talking about stills and beginning video. When you talk about expanding it to other file formats, it’s a heavier lift.

It’s important to stress that this is not limited to Adobe, we’re taking the lead here, but it’s not something we’re trying to sell people, it’s not a monopoly, not at all, it’s an open code that anybody will be able to use, including our competitors.

TTW: What are the problems that could potentially arise? Except, obviously, from your partners not accepting the tool or something along those lines? Is that the main issue?

SL: I think there are some concerns around the security of the metadata and whether it could be hacked or altered. And when we get to the product stage we’ll test the security, we’ll see what we can learn from it. But we’re using advanced cryptographic techniques that are common to the types of security protocols that financial institutions and other large players have. It’s very much state of the art.

I think the other thing is that there are some fundamental misunderstandings out there. There’s an assumption that because Adobe is taking a prominent role here, there’s some suspicion about why Adobe is doing this? You know, is it just to make money? What are we trying to sell people? 

And what we’re saying is, this is an open-source Initiative. It’s not proprietary, we’re making it available to everybody in the same way that for example, PDF is an open-source format, and has become ubiquitous as a document format. Yes, Adobe invented it. Yes, Adobe has very robust tools to process PDFs. But there are scores of other companies out there who will also do the same thing. And we’re not concerned about that.

TTW: Ending on a personal note, why do you do this work with CAI?

SL: It’s kind of the culmination of my life’s work in the sense that I spent many years in the field, as a conflict photographer, often at great personal risk, telling stories and shining lights in very dark and dangerous places. And then I spent 13 years as an executive at the Associated Press tasked with running the world’s largest photo l agency and having a large crew of photographers do similar work around the world. And underpinning all of that work was the notion of trust and veracity and authenticity. So when it comes to being involved with a project like this, which is all about authenticity, veracity and provenance, it’s a tremendous privilege to be working not only for Adobe, but with such a robust coalition of partners, both in the media and technology worlds towards this end.

Responses have been edited for brevity. 

Want to know more about the Content Authenticity Initiative? Check out these resources: 

Introduction to the CAI at Adobe MAX 2019 

Informtional video on CAI

CAI website

CAI white paper

CAI case study

CAI membership information

Tags: ethics, metadata, photography, photojournalism, Photoshop, social media, Visual Storytelling

If you liked this post, check out these:

Balking at free advertising: Why many non-profits aren’t taking full advantage of Google’s offer
Laid off by the LA Times, documentary filmmaker Jessica Q. Chen has more stories to tell
IMG_7202 (1)
Borzou Daragahi explains why writers first need to be visual storytellers