Progress toward a system developed by Adobe to trace the history of digital images, connecting them to their original sources, is steadily advancing. A working model, developed as part of the Content Authenticity Initiative (CAI), is anticipated to be included in the beta version of Photoshop, Adobe’s widely used image editing program.

According to Adobe, a preview of this new functionality will be accessible to users participating in the beta programs for both Photoshop and Behance in the coming weeks. The company characterizes this CAI implementation as an initial iteration of the open standard, which they plan to refine further.

This initiative serves several purposes. A primary goal is to establish a more reliable method for associating creators’ identities with the content they produce. However, a particularly significant application of CAI lies in its potential to function as a “tamper-resistant” industry benchmark, specifically targeting images utilized in the dissemination of false information.

Adobe articulates the project’s objective as a commitment to “enhance online confidence and openness through a comprehensive attribution system that benefits both content creators and viewers.” The outcome is a technological solution that could, in time, curtail the proliferation of deepfakes and other forms of deceptive material found online.

“… Ultimately, you could envision a social media platform or news website offering the ability to filter content based on its likelihood of being genuine,” stated Andy Parsons, Adobe’s CAI director, earlier this year. “However, the CAI deliberately avoids making assessments of truthfulness—our focus is solely on supplying a layer of transparency and verifiable information.”

The concept resembles an evolution of EXIF data, the optional metadata embedded within images that includes details such as the lens used and the location where the image was captured. However, Adobe asserts that this new attribution standard will be more resilient and considerably harder to alter. The resulting system would more closely resemble digital fingerprinting technologies employed to identify instances of child exploitation online than traditional EXIF data.

“We are confident that attribution will foster a positive feedback loop,” explained Allen. “As more creators share content with appropriate attribution, consumers will increasingly seek out and utilize this information when forming opinions, thereby reducing the impact of malicious actors and misleading content.”