According to Wikipedia:
The goal of the C2PA is to define and establish an open, royalty-free industry standard that allows reliable statements about the provenance of digital content, such as its technical origin, its editing history or the identity of the publisher.
Has anyone explored this standard before? I’m curious about privacy implications, whether it’s a truly open standard, whether this will become mandatory (by law or because browsers refuse to display untagged images), and if they plan on preventing people from reverse engineering their camera to learn how to tag AI-generated photos as if they were real.
I know Blockchain is always in search of a solution, but is this one place where it may work? Take a hash of the image and store that hash in a chain, that way you can always hash the image and see if it’s been altered?
Many CSAM detecting services already use image hashing to compare to a central database.
https://www.thorn.org/blog/hashing-detect-child-sex-abuse-imagery/
What value does having a blockchain here provide, exactly?
Publicly traceable and verifiable hashes of the images authenticity. Submitting a hash of the image can prove who submitted it and when and then any altering of the image would yield a different hash which you would know you’re not looking at the original image.
How would that functionally differ from having an authority verify these hashes? Certificate authorities already provide a similar service, and C2PA would likely work in a similar way, sans any effort to implement “trustlessness.”