With generative AI imaging tools getting increasingly sophisticated and all-the-more accessible, it’s getting harder to trust the pictures we see online — but the tech is out there to put an end to the doubt.
C2PA is a picture authentication standard that uses additional metadata to highlight any fakery that’s taken place, but it’s taking years for the key players to get on board and make it happen. @theverge looks at what’s needed to get there.
#GenerativeAI #AI #ArtificialIntelligence #Photography #Editing
@[email protected] @theverge If I understand correctly, the article is referring to the generation metadata which is embedded into images by default? If so there are a number of problems with this. For one, it’s super easy to lose that metadata. Not even necessarily intentionally (most processing removes it.) Even the upload process might even lose it (if any conversion takes place.) Early integration is needed.
Not saying it’s useless, just that it may not be as helpful as this implies.