Towards more reliable press: AFP and its technological bet against deepfakes 📸

Cliquez ici pour lire en français

Agence France-Presse (AFP) has found an innovative way to combat the spread of manipulated visual content. It is implementing a system to certify images and videos, ensuring their authenticity and origin from the moment they are captured.

An unfalsifiable certification solution🛡️

AFP is tackling disinformation with a major technological breakthrough. To guarantee the integrity of its visual content, the agency has developed a process based on the C2PA (Content Authenticity Initiative) technical standard. Moreover, Google, Microsoft, and OpenAI support the project. From the moment an image or video is captured, this standard certifies that it has not been altered. The technology is directly integrated into the camera, where a secure digital signature is generated at the exact moment of capture.

AFP has tested this solution on a large scale using a Nikon prototype camera capable of embedding this signature. This feature is expected to be available on all cameras starting this year, potentially revolutionizing photojournalism practices.

Traceability and security of images 🔍

The certification process relies on a rigorous image storage system. Once captured, photos are sent directly to a secure server where they are duplicated. This ensures the preservation of the original image while allowing modifications according to the agency’s editorial needs. Any modification can be tracked and linked to its original file. This method creates an unfalsifiable certification, making any attempt at manipulation nearly impossible.

Furthermore, an invisible watermark is a key element of this technology. Applied before distribution, it embeds a unique digital fingerprint in each image, linking the edited version to its original file. This indelible marking also applies to videos, allowing for quick verification of content origins.

A new standard for photojournalism🌍

Image source verification is now accessible thanks to the WeVerify plugin, co-developed by AFP. This solution enables fact-checkers, open-source researchers (OSINT), and even the general public to easily decode the invisible watermark. They can then retrieve the original file stored on the secure server. This transparency strengthens journalistic credibility and helps combat disinformation.

In fact, an experiment conducted during the 2024 U.S. elections demonstrated the effectiveness of this technology. Following this success, AFP hopes that other industry players—including camera manufacturers, software publishers, and media distributors—will adopt this standard. The goal is to establish a common certification standard for visual content, fostering greater trust in the information being disseminated.

Towards a secure future for visual information 📅

In the face of the rise of generative AI, AFP’s initiative represents a concrete step in the fight against disinformation. This image certification and traceability system lays the foundation for a new standard of trust in visual information. It could inspire other news agencies and industry players to adopt similar solutions to enhance content authenticity online.

In short, AFP’s innovation highlights the importance of adopting cutting-edge technology to counter the risks of image manipulation. It paves the way for transparent authentication of visual content.

What do you think ??
What are your thoughts on AFP’s initiative to certify image authenticity? Is this an effective solution against disinformation? Let us know in the comments! 😊