P: Marking your photos as not to be used for AI training: c2pa "Do Not Train" assertions

Inquiry 4: The Creator's Right to Choose — Final Clarification
 
"C2PA appears enabled by default across most products, with no visible option to disable it and limited clear disclosure. Please answer:
1. Can I opt out of C2PA data collection entirely while retaining basic authenticity attribution?

2. Can I choose to have credentials applied only to my final exported work, and not to drafts or intermediate editing states?

3. Can I request the permanent deletion of all historical metadata records linked to my account at any time?
If these basic controls are not currently available — is that because it cannot be implemented technically, or because it was never the intention to provide them?"

My question is: who has the right to access C2PA creators' data, creation timeline, and IMEI number (or other data related to mobile phone serial numbers)?Who exactly does this system serve? Are all or only the inspectors? Or is it for some undisclosed "unknown purpose"?

（This is a fundamental question about the technical aspects and future commercial paths for creators, not a forced or critical one. It's a constructive question—this distinction is crucial—and it has nothing to do with emotions.）

 Inquiry 3

Unverifiable Records = An Invisible Archive System
 
"Data is being recorded and stored, yet creators have no unified access portal, no deletion mechanism, no complete inventory of what has been collected, and no access logs. Technically — what is the difference between this system and an invisible creator-behavior archive? If I cannot review what has been recorded about my work, cannot request its removal, and cannot know who has accessed it — upon what foundation does this system claim to build trust?"

What kind of personnel have permission to access the C2PA data center?

Inquiry 2: The Boundary — Recorded Data vs. AI Training
 
"Adobe's terms state: 'We do not use Customer Content to train generative AI models without your explicit consent.' Please confirm: Are C2PA metadata, edit history, device information, and usage behavior data included within the definition of 'Customer Content'? Are these categories of data fully excluded from use in model training, pattern analysis, or feature extraction? If partially excluded — please specify precisely what is excluded, what is not, and where that boundary is documented."

My question is, "Is it true that no one can access C2PA?" or "Who" has access to C2PA's creator data center? (Adobe senior management)It's important to note that this issue has been largely ignored for a whole year since last year. Is it really because it's "unseen" or something else?

I'm daviddesign1988 

 Inquiry 1: Data Scope and Storage
 DAVIDDESIGN1988-
"C2PA publicly states that its purpose is solely to 'indicate provenance and authenticity.' Yet in practice, data collection appears to extend to device identifiers, precise timestamps, edit trajectories, intermediate versions, and linked account information. Could you please clarify:
Which data remains embedded within the file only and travels with it? Which data is synchronized to external servers or cloud repositories? Do Adobe, Microsoft, or other C2PA members maintain a separate, persistent copy of this metadata linked to my works? If so — under what published terms is this collected, for what duration is it retained, and who has access to it?"

2026/10/15

Subject: C2PA “Do Not Train” vs. Data Mining – Are These Technically Enforced?

Hi Adobe team,😄 im DAVIDDESIGN-

I’ve been researching the Do Not Train assertions inside the C2PA/CAI framework (like c2pa.ai_training and c2pa.ai_generative_training).
From what I understand, these seem to act as declarative preferences rather than enforceable restrictions.

However, since many AI datasets are built through data mining or feature extraction processes — not explicit “training” —

I’d like to ask:

1️⃣ Are these C2PA “Do Not Train” assertions technically enforceable at any level of data ingestion?

2️⃣ Does Adobe plan to extend the standard to cover data mining / embedding / feature analysis use cases (e.g. c2pa.mining_training)?

3️⃣ If a third-party crawler strips metadata before indexing, does Adobe still treat the original C2PA declaration as legally binding in provenance verification?

These seem important questions, since without a cross-layer enforcement mechanism (between metadata and dataset parsing),

the “Do Not Train” flag may remain symbolic rather than functional.

Thanks for clarifying — I believe many creators would appreciate a technically transparent answer here...

https://community.adobe.com/t5/photoshop-ecosystem-ideas/add-support-for-iptc-plus-coalition-generative-ai-ml-training-metadata/idi-p/14808232