Starting with personal anecdotes from an extraordinary career, Lyon reminded the room of the high stakes involved. “I spent 20 years in the field as a photojournalist, covering nine wars on four continents. Then I went to being Director of Photography at Associated Press (AP) for almost 15 years, overseeing about 1000 photographers worldwide generating about a million images a year - all of them truthful and honest representations of the world,” he said.
Lyon now focuses his efforts at Adobe on tackling an increasingly confusing media landscape, where generative AI - which has been the great disruptor of the past two years - signifies a double-edged sword. “When I saw that AI-generated image of Pope Francis in the white puffer jacket, I thought it might be real for a moment. Why not? A progressive Pope in Vatican City, next to Rome, the fashion capital, seemed plausible. But I was wrong,” he admitted. “If I can be tricked, I can only imagine what a less informed viewer might think.”
These missteps, he noted, are part of a broader problem. “Gen AI has cast a shadow of doubt over everything. It's what some people refer to as the liar's dividend. If I can lie about something, I cast a shadow of doubt over everything,” Lyon pointed out. With generative AI blurring fact and fiction, the veteran photojournalist highlighted the importance of establishing trust through transparency, a principle that drives the CAI.
Founded by Adobe in late 2019, CAI has grown into a coalition of over 4,000 members, including companies like Reuters, AFP, Microsoft, Getty Images, Nikon, and more. The initiative is focused on providing content provenance, enabling creators, audiences, and publishers to verify the origins and edits of digital files. Lyon described the CAI’s mission as “proving what things are, instead of simply detecting what’s false.”
“The main focus here is provenance – we’re talking about the basic facts about the origins of a piece of content, where did it come from? How might it have been manipulated along its journey from creation through editing and onto publication, and then sharing some or all of that information with the viewer so that they can make a better informed decision.”
CR - A new icon of trust
Lyon described one of CAI’s most innovative tools - content credentials (CR), which provides details like whether an image came from Adobe’s generative AI tool Firefly, or had been edited in Photoshop. “So provenance then really becomes about proving what things are, as opposed to detecting what's false,” said Lyon.
The way this is done is by sharing relevant information through the “CR” icon, which signals that additional details are linked to the content. Users can interact with the icon to access and explore this information. Through partnerships with camera brands like Leica; and even Gen AI providers like OpenAI and Adobe Firefly, the CAI is embedding content credentials into the DNA of digital media. TikTok also joined CAI and C2PA (Coalition for Content Provenance and Authenticity).
“Time Magazine gave us a little pat on the back, saying this was one of the best implementations last year,” Lyon proudly noted.
One of CAI’s standout features is its deliberate, inclusive and collaborative approach. Everything the CAI does is open-source, which means Adobe’s competitors can use it, to encourage transparency and authenticity.
Organised by the Sharjah Government Media Bureau (SGMB) Xposure 2025 is taking place in Aljada, Sharjah until February 26. To plan your visit, see
https://xposure.net/.
For an optimal experience, please
