Content that most needs verifiable provenance - content shared virally on social media - is precisely the content most likely to lose its C2PA metadata during distribution. Open to solutions...
Society needs both. I'm not arguing otherwise.
But I'm thinking about models for sustaining journalism. And that question starts at the point of capture.
In the AI provenance question, does it make more sense for news publishers to label and identify AI-generated content - or track the provenance of captured footage and recordings?
My feeling is the latter. Here's why.
Verified capture is a more fundamental solution - with direct value for publishers - from which broader policy can work outward.
Captured footage provenance isn't a subset of the provenance challenge. It's a more fundamental pillar.
Canon just launched a camera system for news organisations that cryptographically signs images at the moment of capture. Not in post. At capture. I like that.
The provenance record is embedded before a single edit happens. It travels with the file. It can be verified by anyone.
That's not a feature. That's infrastructure.
And it will help ensure preservation of culture.
South Africa is in the EMEA rollout. Whether any newsroom here is paying attention is a different question.
To all the news publishers out there - whether you like it or not - chatbots, whether on Claude, ChatGPT, Gemini...are becoming platforms. Open up your models faster and start listening to guys like me